UiPath Aktie

Good morning and thank you for joining us for the UiPath session at the Growth Stock Conference. I'm Pat McIlwee, and I'm an analyst here in the software group at William Blair as a part of which I cover UiPath.

I'm required to inform you that a complete list of disclosures and potential conflicts of interest are available at our website, williamblair.com. We're very happy to have the UiPath team back at our conference this year, including COO and CFO, Ashim Gupta, as well as Allise and I think, Jake as well from the IR team here in the audience.

UiPath is one of the leading automation and orchestration software providers for the enterprise, enabling businesses to automate repetitive digital workflows across applications, systems and business functions. The company just reported their first quarter results last Thursday, and their positioning to benefit from the adoption of AI and the enterprise is increasingly clear. So I think it's a great time to be digging into this story. So that's my quick two-liner on the business, and I'll turn it over to Ashim for a better overview of the company.

Awesome. Thank you, everybody. So I'll put the safe harbor up for a second. It's actually really great to be a part of this team and to see everybody here today. I would say we came off of one of -- like, to me, one of the most foundational quarters that we've had. And the reason why I look at it is there's 2 parts of a quarter. It's kind of what you deliver and the foundation about -- that foundation that you're building to deliver kind of the next stage of growth for the company. And for UiPath right now, if you look at where we are, go back to founding of UiPath, 5 years, we took RPA and made it a $1 billion revenue business. In 2019, we branched out from RPA. And I -- if there is kind of a couple of bullets I would impress is like we are not an RPA company. RPA is one part of our platform. But in 2019, we began adding to our AI portfolio, both IDP in terms as a intelligent document processing for structured and unstructured data, process intelligence, communications mining. When you look at where we are today with 18 months ago, launching our agentic -- our business process orchestration, which is Maestro that orchestrates humans, robots and agents. Our AI ARR now is $200 million. That is not switching a metric, changing a definition. That is solid ARR that is coming from real value that we are providing customers with our AI platform, and look forward to talking more about it in our breakout sessions and within our Q&A.

Then you look at fundamentally where we are. $418 million of revenue, that grew 17%. Our ARR growth rate at $1.9 billion is 12%. And when you look at that, that is now stabilized. I think one of the big questions for UiPath was the trajectory. And when you look at last 2 years, we had decelerating ARR and revenue, really some execution missteps, but also some of the change in macroeconomic conditions, et cetera. Right now, if you look at our last couple of quarters, especially solidified in this quarter, we have stabilized the business and the growth rate at greater than 10% as a company. And that is kind of on the backdrop of what continues to be a pretty macroeconomic variable environment. The other big important thing is we're not spending money to get growth. We're not buying growth necessarily. So we are GAAP profitable now. And that is our first, first quarter of GAAP profitability. Again, you look at trajectories, 2 years, we had fourth quarter where we were GAAP profitable. This is the first quarter that we are profitable for our first quarter. And you can just start tracing the trajectory of our stock-based compensation going from greater than 23%, down below 13% or at or below 13%. And so when you look at that together, foundationally, we have a strong AI platform, solid growth at scale, but which we're not fully satisfied with and we'll talk about it further, and a really good profitability equation, both on a GAAP and non-GAAP basis.

So here's our platform. When I say we're not an RPA company, you can see RPA is one box there. I don't want to discount RPA. If I called RPA tokenless automation, or if I call it, deterministic automation. That really resonates in terms of the applicability that it is with the customer. So this isn't basic screen scraping. The deterministic parts of our platform are powering significant enterprise grades automations in highly regulated industries. But look what's now built around it. So we are one of the few platforms that can have human in the loop, build deterministic automations and deploy them on our platform, build and deploy AI-based automations or agenetic automations on our platform. We are not vendor locked in, so people can build -- bring their own models, their own agents to our platform. Our goal is to orchestrate the workflows for regulated industries and complex enterprise processes globally, that ties robots, agents and humans together. So one other misnomer that you may hear, agenetic orchestration is different than business process orchestration. Agentic orchestration is orchestrating agents. Agents are one piece of the equation if you want to drive enterprise automation at scale. These are our advantages, and we'll touch base on them more, but we are a truly unified platform for what I just talked about. What is exciting is we are an incumbent. So we are not trying to go -- if you think about the hundreds and thousands of automations that are there on a deterministic side, one of the ways we were able to grow our ARR base is by looking to the left and to the right and being able to get agenetic processes or agenetic steps that were not previously automatable and then wrapping that around orchestration. And then besides governance, which is something really critical, we'll talk about in Q&A, we have deep vertical expertise, health care, financial services. It's not enough to have software and code. You have to understand the processes and that is an expertise we've been building over the last 5 to 6 years. So these are our customer metrics. You can see our largest customers are growing. We are super proud if there is a metric that is there to look at. It is our customers greater than $100,000 and our greater -- customers greater than $1 million, you can see that customers who know us, these are Global 2000 customers as the majority of them, they are growing fast with our platform. That should show you that our platform is not just relevant today, it is value-added.

And then last is the partner ecosystem. We'll talk about it, but we value this not just because it's a page that is important to all software providers, but it really speaks to the openness of our architecture and how we can partner across every different cloud base that is out there. Every different technology company that is there. We're not married to anyone. So with that, I'm really excited to answer questions. And I thank you for everybody, and I look forward to your questions as well.

Awesome. Thank you, Ashim, for providing that perspective on the UiPath business and the toolkit that you provide to these businesses. With that context, can you just set the table by kind of in simplistic terms, talking about the distinction between deterministic automation and then more probabilistic automation?

So deterministic automation is rules-based automation. So where you need the same answer every single time. Probabilistic or agenetic orchestration is giving software the agency to make decisions. And as it does that, it will make a different decision every time. I'll give 2 quick examples. And for time, I'll try to be brief. Daniel, our founder actually gave me this challenge when I -- when we were talking about it. Go into Claude or OpenAI and ask it to calculate 2 large numbers. It will rely back on a calculator because that is deterministic. But if you say, do not use a calculator, will you be able to produce the same answer every time? It's answer, actually response is, no. Because LLMs or the models that are there, they're probabilistic, they interpret data each time differently. When you look at processes like claims, mortgage, accounts, I shouldn't -- my Chief Accounting Officer and Deputy CFOs in the audience, we don't want to have 2 different calculations for revenue depending on the mood of a model, right? You need deterministic automation. If you have a healthcare claim, you need to be able to look at it. So deterministic automation is super important for enterprise-grade processes because it is low cost, low complexity and the highest level of dependability. Agenetic has its place. There are things that deterministic automation can't do. So when you put the 2 together and then you wrap it with the ability to orchestrate it, you have the chance finally to get to a fully automated enterprise, which we are super excited to be a part of.

That's great. And so you guys obviously provide a platform that encompasses both of those capabilities, deterministic and probabilistic and there's some harmony between the 2, right? But historically, you've been kind of the undisputed leader in RPA market using bots to automate repetitive tasks. So as AI continues to make its way into the enterprise, how do you see those RPA workflows evolving alongside this technology?

They're very synergistic. When you look at our customer base, no one really sees a bit, if -- no one has come to us to say, we are going to take a process that is working that is low cost, that is highly dependable. And we're deciding to move it to a higher cost, lower dependability solution, right? The area -- the only area where we see -- where we have seen that piece of it ever come in is personal productivity. And I think one of the misnomers for UiPath. In 2019, we launched something called a Robot for Everyone, right? That is a very low part of our base. That means if I want to go and download an e-mail or summarize a document, right? Those are things our platform can do today. And it was part of a strategy back in 2019 -- 2018 and '19 in terms of where we are. When you look at our business today, it is regulated industries and it is high complex enterprise grade automation. That is the majority of our revenue. So when you look at from where we stand today, you need deterministic and agenetic to really drive that in that tier of relationships.

Got it. And before you move on, given we have a lot of journalists here, can you just talk about your pricing model, how the digital bots are priced, how your orchestration solution Maestro is priced and how you present the ROI to your customers in your sales motion?

Yes. So one thing that's like a great advantage of UiPath is we're not really, [ seat-based ] pricing is not a major component of our pricing. So let me tell you what it's not to start with. The second piece is we don't have tokenization at this moment. So there's no like large-scale token consumption that is driving a short-term revenue boost in any which way. The way that we price is server-based pricing for our key deterministic parts of our platform like unattended robots. We have what I would call kind of subscription consumption-based pricing, which is you buy a certain amount of units, right? You buy $1 million worth of units. And for every page, you deduct [ X units ], right, as an example, that you process. For every execution that you do on our orchestration, you can retire x units that are there. Those are the 2 primary methods of our pricing.

Got it. Okay. And I think it's clear, we'd kind of be beaten around the bush if we didn't acknowledge there's some fear of disruption associated with some of this technology. But when speaking to your customers and customers of your peers alike, it seems like they're leaning more into your trusted platforms than they are trying to move away from them in this environment and as they execute on their AI strategies. Can you just talk about that dynamic and what your customer conversations look like at this point?

Yes. I mean, look, put yourself in the seat of any kind of company executive. You have this wave of AI that is going with these really loud and important voices, OpenAI, Anthropic, right, Google, et cetera. So I think there's a mindshare that is being taken up right now that we have to acknowledge at the corporate levels, trying to explore what's possible in AI, right? That creates some level of disruption, that creates some level of fear, that creates some level of uncertainty of what is there. The way we -- the way our customers have responded is usually what's happened is as they process, where do they really want to deploy these models? What's the real impact of it? There is a place for those companies, and there is a place for UiPath. And UiPath is actually a great channel for a lot of the models that come in, right? In terms of pure-play competitors, we don't really see. There's not -- if you go and say, which companies have RPA, unstructured document capabilities, processing capabilities, agentic capabilities and business process orchestration, not agenetic orchestration, there are very few companies that do it. So really, the discussion with our companies is really targeted. And so we're selling more and more into lines of businesses with specific outcomes in mind, and that is really helping us. And the second piece is we verticalized. I think one of the things that people have not seen, and I think we can do a better job showing it is, we used to be purely a horizontal platform. Now we have products and capabilities for revenue cycle management and healthcare. We have processes for procure-to-pay, software for procure-to-pay in the office of the CFO. And I think that is also driving further differentiation for us.

Okay. Yes, that's great and very clear. So I think it's a good segue into my next question. Maestro is your control plane, your orchestration plane for agentic processes, which makes a ton of sense, given how embedded you already are in these workflows. How do you see Maestro competing with competing solutions from other platform players, other AI management solutions? And do you ever or at all see frontier model providers as competition or more so as partners?

We see it more as partners. LangChain is an example. We have actually a great integration, great partnership with them, just to give an example of it. From our standpoint, the differentiation is, one is trying to manage AI, and we are trying to manage processes, regardless of the if there's an AI piece of software embedded or an AI automation in there or deterministic automation or a human or 2 or 3 of them. And that is really our key main difference. The other piece is observability. So what I would ask everybody to do is Google Maestro UiPath. When you do it, look at the click on images, right? You don't have to read all of the technical documentation. You can, if you wish. But look at the image, you actually will see the process. And if you click on video, you'll actually go and be able to see the process moving. So transactions move through the process. Imagine a world where you have digital workers, that you can now see as though it is a factory floor. That is for us, that is kind of how we think about Maestro. No, we do not think about the frontier models as competition. We think about them as integration points that we have to continue to drive within -- with our customers and within our partner ecosystem.

Okay. Great. And on pricing, so as you embed AI into these RPA workflows, into your broader platform, you win the orchestration layer, it becomes capable of executing more complicated multi-step processes, right? So what is the typical pricing uplift see when you sell those solutions? And is there a point in the future when those kind of become table stakes?

Two ways I can answer it. One is like our pricing scheme, I think, is dissimilar to what we talked about. We are experimenting with outcome-based pricing that provides significant uplift, but we take some upfront risk as a part of that. There are several customers that we're exploring that with. If you put that aside, if you look at first quarter, 16 of the 20 deals were based on, you had AI components as a significant component as a part of the platform. They were 6x larger than all of our other deals. When you look at process -- when you look at our $1 million-plus customers, the majority of them have large attach rates for AI. So from our standpoint, the more value we generate, the more software gets pulled through, the less discount that comes in, those all go to higher and higher ticket sizes. And where you see it show up is our $1 million-plus customer accounts and our $100,000-plus customer accounts, which are up 11% and 16%, respectively.

Okay. And then kind of to shift gears. You've also talked about how these coding agents can lower the time to value, the implementation time for these workflows, ultimately allowing you to go after more of the long tail of these automation opportunities in the enterprise. So can you talk us through that dynamic and how significant that reduction in time to value can be?

Yes. You're talking from months to weeks or from weeks to days. It is really significant. And from our standpoint, like the best metaphor, our CPTO, Raghu Malpani give is, our goal is that you can 3D print processes, 3D print workflows. And why I think that matters is if you thought about like today, the experience of building an app, on Lovable, on Replit, on some of these platforms, which are incredible. Imagine being able to print a process in the same way. And then as you do that, the question I get is, hey, why don't you do that on those other platforms? Well, think about it this way, like if you've made an app on one of these platforms, you know the difference between an app and a process, right, what it can do. And so from our standpoint, from week -- from months to weeks, from weeks to days, that's the goal. And we've already launched our first set of coded agents. This isn't hyperbole. The results are actually super promising in terms of what we're seeing with our early customers.

Okay. Great. And I think 6 months ago, you launched a forward deployed engineering team. So can you talk about how that ties into that strategy?

Yes. So the piece we didn't talk about for time to value is, I think there's 2 time to value equation. One is how do you make development on your horizontal platform faster. The second is when you go across an industry like health care, most every health care provider that we see is looking for certain solutions that they have been automating, prior authorization, claims denials, revenue cycle management. When you across -- go across the financial services industries, mortgage processing, HELOC filings, et cetera, right? Our verticalization, our FTEs have twofold areas. One is to be working with our customers so we can enhance our product to get greater verticalization into our product as we launch our vertical solutions to be there to perfect them, to add features and functionality that's there. The second is we use our FTEs as we deploy the first wave of agents. Many of our POCs were bolstered by FTEs for 2 purposes: ensuring success; and the second piece is getting the feedback back to our product. That's kind of in the early wave of where we are. As we go forward, we really think FTEs are going to become more and more a tip of the spear type area for our key customers, because this place is going to be very fast moving. The demands are going to be higher, the complexity is going to be higher, and they really serve as that connective tissue between what was a services organization and the product organization.

Got it. Okay. And does that have any margin implications? I know it's early on.

No. Actually, what is -- so investing in FTEs alone, of course, as we add headcount. However, the way our approach has been, we are investing in 4 deployed engineers, key R&D areas like coded agents and vertical solutions and sales capacity. These are kind of 3 key areas that we're investing in. We are divesting in every process that we can agentify ourselves. So when you look at our margin, I feel like this is the area that's there. We're not in cost-cutting mode. We are in a really strategic capital allocation mode. So when you look at whether it's geographic territory, whether you look at management layers, whether you look at centralized organizations, we are constantly driving efficiency into those areas, and we are increasing our capacity and our capability into the areas I just said.

Okay. Got it. And I'm going to jump to one of my last questions actually because it's on the same topic. But you made some really solid progress on your margin on the bottom line over the last few years. And given that progress, you raised your long-term operating margin target to 30%. Can you walk us through what levers you have in mind as you think about making it to those targets? Or if it's more of kind of a North Star over the [indiscernible].

No, I think -- look, I think every quarter, every year, I feel for the last 2 years, we've significantly improved our margin. I personally think free cash flow margin and GAAP profitability are the 2 most important things. The 2 most important things I look at. So if you look at our free cash flow, like last year, we were at $370 million. The year before that was significantly lower. This year, we've kind of given a modeling point of around $425 million in terms of our free cash flow, right? So I don't -- the first thing is it is a North Star, but it's a North Star that is reachable. It's not something that we're just directionally going after. The levers we have are 3. One is, frankly, cut nonsense out, be a highly efficient, be highly focused and prioritize its basic capital allocation, prioritize your -- both your human and your expenditures where you have the highest return, that's the first principle. The second area is as we do that, as we identify processes internally while investing. If we can keep our cost base relatively flattish while growing top line, you naturally get there very quickly. But we are not afraid to invest to be able to grow. We actually feel like right now, we're kind of in F3 of where we are as a company, F2, F3, we are going to invest to make sure that we can take a part -- take a large part of the market that we have in front of us. So I would say long-term goal, 30% where I think 23%, 24% already, 6 points of operating margin between leverage and discipline completely within our grasp. And then the only question is, if we see market opportunity, we'll pace that accordingly.

Okay. And is there a portion of that where you're leveraging AI to drive operational efficiencies as well? And to use the term, I have recently kind of eating your own cooking.

Yes, yes, drinking your own champagne, eating your own cooking, whatever it could be. Eating your own dog food has been like -- has been put away. Hitesh Ramani, who is our Deputy CFO. He's actually in the audience. He's leading that effort across a lot of our back-office functions. We have 70-plus agents, I think, in production already. And we still feel like we're just at the starting point, right? We were able to significantly drive transactional efficiency. We're able to transform functions like marketing, we're able to transform and get productivity out of engineering. So in the past, if you look at our product road map, I would say our product road map is probably 3x the surface area, but it is with the same number of people. So that is coming from coded agents and using the tools even in the engineering shop in terms of the productivity they're getting. So we are kind of -- we're doing that. And within our own platform, we continue to drive automation. We continue to drive agents. And the orchestration and differentiator in terms of doing it now larger at scale, but still being able to have the right controls in place being a public company.

Got it. Yes, that's very interesting. Thank you for the color on that. And something that you and Daniel have talked more about recently that I think is very interesting is the test cloud opportunity as these agentic workflows proliferate. Can you just talk through exactly what that product is for the audience and how large that [indiscernible]?

Yes. So I'll -- test is like -- it's a dark horse for us. I think I purposely also let it be a dark horse for a bit, but you're going to see us kind of like more and more emphasize it. If you just take what application testing is, everybody, I think, should be pretty familiar with it, right? You have upgrades into your software and then you have to have scripts to make sure before you put that software into production, the key transactions are flowing. So about 4 years ago, it was a very logical thing is that if you're automating the actual process, why are we not automating in the scripts that have to be written to ensure that the applications are in good standing as they get updated and moving. So if you think about -- take every S/4HANA implementation that's happening, you have a massive amount of transactional testing that has to be done. A lot of that is manual or a lot of the software in this space is archaic. We actually have the most modern platform now. It has been recognized as a leader by Gartner in terms of the category in which it's in. What's exciting is that it opens up a different buyer base within our company. So we're now selling into true CIO organizations in terms of QA/QC, et cetera. And we've seen -- we haven't disclosed it at this time, but that -- the ARR growth on that has been super exciting. It's kind of operating like a startup within a startup and is a really second throttle of growth for us as a company.

Got it Okay. Yes, very interesting, but we'll leave it at dark horse for now. So we have a few minutes left. Anyone in the audience want to jump in with a question. Otherwise, I can ask a couple more to finish this up. There will be plenty of time to breakout for questions as well.

Okay. So Ashim, you've continued to maintain best-in-class gross retention rates, high 90% range. But your net retention over the last couple of years has dipped a little bit until the most recent quarter, right, when we saw it step up 2 points to 109%. So as we think about everything we've just discussed, can you talk about what's driving that inflection? How you all are driving customer expansions and just walk through the mechanics of those expansions a little bit?

Yes. I think, I would say 50% to 60% of it is just better execution, which Daniel has really driven over the last 2 years with the -- getting the field flatter, getting it close -- getting management closer to the field, driving more disciplined motions. So we have lost touch, I think, 2 years ago with our customer base. And we openly talked to that. I think what you're starting to see now is kind of the rebuilding of both trust and relationships and that intimacy with the customer base that is, I would say, a foundational ingredient of driving more and more software and more and more processes through our platform. The second piece is we have more products to cross-sell. So when you look at Maestro, when you look at IXP, which is advanced unstructured document processing, when you look at WorkFusion, which we acquired, Peak that we acquired, you have more and more parts of our platform that you can sell across our installed base. So our installed base is super powerful. And as we get more products, as we're closer to the cutomer, that is -- that gives us a real chance to drive that leverage. And what we're cross-selling in is both deterministic as well as AI. If you look at leading segments, healthcare. Healthcare, it is all AI all the time, and it pulls forward, pulls through deterministic automation that is needed. But if you look at the public sector within Europe, RPA is the hottest topic. So depending on the market, we're able to continue to drive upsell across our platform. And what's exciting is, I think U.S. healthcare is actually a pretty bleeding edge type of industry, when it comes to technology, and it's [indiscernible]. So now I have a lot of confidence that says in the European market that same dynamic can start coming.

Okay. Great. We've got 1 more minute. So I'll sneak 1 more in. So you talked about the kind of credit base, retire the credits over time as you execute on these workflows. How -- is there any way you'd expect that to change over time as you continue to iterate the platform?

The answer is, I don't know. Just in all candor. I think we have a monthly pricing council. I think the world is evolving very fast. What we don't want to do is like knee-jerk react to change our pricing model to like jump on a bandwagon, so to speak. Our best and most reliable source is our customers themselves. So what customers are really nervous about today is like just unguarded token consumption, token burn. You've seen public companies come out about it. You've seen different posts and different narratives being there. So our advantage point is like we want to give predictability of cost to our customers. I think if you have unpredictability of cost, it hampers a long-term adoption. It may feel short term, but it really hampers long-term adoption. So from our standpoint, predictability is going to be a key pillar, and then we'll see how the industry evolves.

Okay. That's great. Well, thank you so much, Ashim for being here. Thank you all for attending the UiPath presentation. And I believe our breakout is in [indiscernible] upstairs. So we'll see you there in 10 minutes.

Thanks.

good day, everyone. My name is Megan, and I will be your conference operator today. At this time, I would like to welcome you to the UiPath's First Quarter 2027 Earnings Conference Call. [Operator Instructions]

At this time, I would like to turn the call over to Allise Furlani, Vice President of Investor Relations.

Good afternoon, and thank you for joining us today to review UiPath's First Quarter Fiscal 2026 Financial Results, which we announced in our earnings press release issued after the close of the market today. On the call with Daniel Dines, Founder and Chief Executive Officer; and Ashim Gupta, Chief Operating and Financial Officer, to deliver our prepared comments and answer questions.

Our earnings press release and financial supplemental materials are posted on the UiPath Investor Relations website. These materials include GAAP to non-GAAP reconciliations. We will be discussing non-GAAP metrics on today's call. This afternoon's call includes forward-looking statements regarding our financial guidance for the second quarter and full year fiscal 2027 and our ability to drive and accelerate future growth and operational efficiency and grow our platform, product offerings and market opportunity.

Actual results may differ materially from those expressed in the forward-looking statements due to many factors, and therefore, investors should not place undue reliance on these statements. For a discussion of the material risks and uncertainties that could affect our actual results, please refer to our annual report on Form 10-K for the year ended January 31, 2026, and our subsequent reports filed with the SEC.

Forward-looking statements made on this call reflect our views as of today. We undertake no obligation to update them. I would like to highlight that this webcast is being accompanied by slides. We will post the slides and a copy of our prepared remarks to our Investor Relations website and immediately following the concluded on this call. In addition, please note that all comparisons are year-over-year unless could have otherwise indicated.

Now I'd like to hand the call over to Daniel.

Thank you, Allise. Good afternoon, everyone. Thanks for joining us. We delivered a strong start to fiscal 2027 once again exceeding our guidance across all key financial metrics. Before I dive into the results, I want to take a moment to reflect on our progress over the last year, In May of last year, we launched our agenting and business process orchestration products into general availability. One year in, adoption has moved from or experimentation to production deployment. .

We are seeing this play out across 3 areas in particular: installed base expansion, process orchestration adoption and vertical workflows. A great example is one of the largest health care distribution companies in the U.S. One end-to-end workflow, combining UIPath agents and deterministic automation is expected to drive multimillion-dollar annual savings, which led to a 7-figure expansion in the quarter.

One of the world's largest construction companies adopted our purchase-to-pay vertical solution and told us they chose you. Not as software vendor, but as a strategic codevelopment partner for their enterprise AI transformation. And the Fortune 500 energy company placed UiPath, but the center of $70 million cost reduction initiative made possible by our ability to bring deterministic agentic and process orchestration together as a single plus.

Turning to the quarter. First quarter ARR reached $1.901 billion, up 12% year-over-year, driven by $49 million of net new ARR and revenue of $418 million, up 17% year-over-year. We grew first quarter non-GAAP operating income to $92 million a 22% margin, driven by improved operational efficiency and disciplined execution across the business. We delivered first quarter GAAP profitability for the first time in company history.

This quarter's performance is built on the strength of our enterprise automation installed base thousands of customers with deep platform adoption, proven ROI and a track record of expanding with us over time. and it reflects continued momentum with our AI products. In the quarter, 16 out of top 20 deals including AI and expansion deals that included AI were larger than those that did not.

The drivers behind these results are the same core differentiators we outlined last quarter our platform that brings together deterministic and agentive automation with enterprise-grade process orchestration, our installed base slideware, our governance foundation and our ability to combine a horizontal automation platform with deep vertical solutions.

I saw that momentum firsthand across our global events including in India at our annual fusion event and that can developer conference. Across customers, developers and partners. The message was consistent. Enterprises increasingly need the platform that can cover and orchestrate humans, agents, workflows, automations and systems, an area where UiPath offers a structural advantage. At Devon, we launched UiPath for coding agents, enabling developers to connect their holding agent of choice. To create, test, deploy and manage automations across the full life cycle on the UP platform with enterprise-grade governance and reliability built in.

This matters because nearly every customer conversation surfaces the same constraint, and automation backlog that outpaces their capacity to build and maintain Implementation is often harder part, particularly in complex enterprise environment, where upstream system changes can drive maintenance costs over time. By combining coding agents with the governance, orchestration and self-healing capabilities built into our platform, we can dramatically reduce the operational burden and compress deployment time less from quarters 2 weeks.

We expect this to accelerate time to value for our customers. drive deeper adoption and strengthen long-term retention across our customer base. Our internal teams and customers are also seeing great results with coding agents, including 1 of the world's largest consumer electronics companies, which reduced a 4-week project built to 3 hours and one of the world's largest chip manufacturers reduced a 2-month project build to a few days. What stood out most this quarter is how clearly customer priorities before with the focus consistently centered on process orchestration, as one customer put it during that content, models are easy. Orchestration is not that directly reflects what we hear across our customer base.

Customers are no longer asking us simply to deploy more agents or generate more code. They are asking us to transform our entire business functions operate through end-to-end workflows that span departments, connect systems and deliver measurable operational outcomes. And delivering that kind of transformation requires more than individual AI agents. It requires a platform that can orchestrate agents, automation, API systems and people together within secure government enterprise workforce.

A great example is one of the world's largest telecommunications companies with nearly 2,000 processes already automated and all the $30 million in annual cost savings, they are now expanding their that terministicbase further and moving into agent workflow building a pipeline of more than 200 additional deterministic automations and over 20 agenetic use cases. That same process orchestration capability also drove a competitive displacement with the Fortune Global 500 electronics manufacturer where we were the only platform that could take them from task-based automation to enterprise-wide business process orchestration.

Building on a strong deter listing foundation, they are now expanding across manufacturing and supply chain workflows using Maestro to coordinate automation, agent, systems and human decisioning globally. Maestro already excels at structured workflows like invoice approvals and deployment pipelines where the process itself is clearly defined. But increasingly, enterprise work is nonlinear and its dynamic, exception driven and center around decisions that move across teams and systems. This is why at Devon, we launched Maestro case into public purview extending beyond traditional process orchestration into the orchestration of unstructured enterprise work.

The breadth is what makes a the most complete process orchestration and automation platform in the market, and it's already driving broader customer adoption, including Sonic Automotive, an early adopter of our agenetic product. They initially deployed new aircraft to automate vehicle stocking and sales lead all up. They are now standardizing their Agentic automation strategy on the UiPath platform under a broader C-suite initiative and expanding into workflow such month and close and employee onboarding. The key driver of the expansion was Maestro case's ability to orchestrate complex multistage workflows across agents, automations and people.

Beyond process orchestration, documents remain 1 of the biggest sources of friction in enterprise world. And customers are increasingly turning to UIPath XP to automate document-intensive workflows at enterprise scale. In May, we were named the leader in the Forrester wave document mining and analytics platforms Q2 2026. We are seeing that momentum translate directly into largest enterprise deployments and competitive wins A great example is the leading medical technology company that is standardizing on UiPath XP to automate high-volume and structural documents like invoices and purchase orders.

The customer is already realizing approximately $5 million in annual savings and expect that to grow to $10 million as they scale. For industry-specific government workflows continues to grow as enterprises increasingly adopt purpose-built AI solution tailored to their business. What differentiates UiPath is our ability to combine them with deep domain-specific solutions with the same process orchestration, automation and governance platform. This quarter, we expanded our portfolio across financial services, retail and manufacturing and the office of the CFO.

We are already seeing momentum in health care in a 7-figure new logo win, a leading Latin American health care provider, selected our vertical solutions to support revenue cycle management, medical record summarization and claim denier management and expect $12 million in cumulative benefits. Customers are also realizing meaningful operational benefits from these vertical solutions, a leading health care technology company produced clinical summary review times by 90% using our medical record monetization solution.

We are seeing similar momentum in financial services a digital bank is now automating 61% of sanctions heat reviews with our transaction screening alert review solution, processing roughly 14,000 or less per month. and is accelerating software creation, but is also accelerating the need to validate it as coal volume grows, so does the testing bore, Independent research for have consistently recognized UiPath as a leader in this space, and we believe that validation reflects the real and growing structural abandon.

This cloud is at the center of that, helping customers move testing from a downstream bottleneck to a continuous intelligent function embedded across the delivery life cycle. One example of this quarter is the leading U.S. utility provider that adopted UiPath Test Cloud for gentic testing to streamline customer platform support launch. The solution is expected to significantly reduce manual testing while generating nearly $3 million in savings. During the quarter, We continue to deepen our partnerships across both go-to-market and technical integrations.

This included our expanded collaborations with Deloitte. Embedding UiPath, Test Cloud into the ascendelivery platform, bringing agentic testing capabilities to Deloitte's global client base. We are seeing similar momentum with Accenture, a life sciences customer, we highlighted last quarter, worked with Accenture to deploy global genic sales entry solution and has now came based across 70 countries. Building on that success, they signed a seve-finger expansion and are now partnering with us to design an office of the CIO intake solution built on our process orchestration platform.

On the technical side, we continue to broaden our reach across key enterprise ecosystem. With Microsoft, we integrated UiPath security this to help automate trade detection and response with Salesforce, we launched a new agent exchange offering that extend my extra process orchestration across sales force and back office systems with Google Cloud, we brought our ISP solution to their marketplace. And with data bricks, we connected their data intelligence platform directly with IPO process orchestration to help enterprises move from data insights to automated action we then govern workflows.

In summary, this quarter reflected disciplined execution across the business, continued AI adoption and growing momentum across our platform. No other vendor can bring together deterministic automation, agent AI, document intelligence and business process orchestration on a single platform. And that complexness is what customers are standardizing on. We believe we are uniquely positioned for this next space of enterprise AI adoption and our strong start to fiscal 2027 reinforces both the durability of our business and the scale of the opportunity ahead.

Before I turn it over to Asim, I want to take a moment to acknowledge the loss of our dear friend and board member, -- so massage. Soma was the long-time investor in UC and we joined our Board just 8 months ago. its impact on UiPath was immediate and profound. He was the mentor drastic adviser and someone I deeply admire both professionally and personally. I will miss him greatly -- and I know our entire Board and leadership team share that feeling. Our hearts are with his spending.

With that, I'll turn the call over to Ashim.

Thank you, Daniel, and good afternoon, everyone. Before turning to the financials, I'd like to provide a quick operational update. We continue to make meaningful progress across the key priorities we outlined last year. Our partner ecosystem is becoming more deeply integrated with both our go-to-market motion and customer adoption efforts, helping us scale larger enterprise deployments across industries. As Daniel mentioned, partners like Deloitte and Accenture are increasingly instrumental, not just in selling, but in helping customers operationalize and scale AI-driven work. and we are seeing that play out across financial services, health care and other key verticals.

At the same time, our internal focus on customer adoption remains a central operating prior. We continue to invest in our services organization and industry expertise to help customers accelerate deployment and expand platform usage. A key part of that effort is our forward deployed engineering program, which we launched 6 months ago, are proving to be an effective bridge between product innovation and customer deployment, shaping vertical workflows directly in customer environments and accelerating time to value. In addition to adoption, our go-to-market teams are executing with discipline and customer interest. AI is now part of virtually every strategic customer conversation.

And those discussions are increasingly expanded into platform, orchestration and vertical solutions. The deal data Daniel mentioned reflects that. AI was included in 16 of our top 20 deals and expansion deals that include AI were 6x larger than those that did not. Finally, on operational efficiency, AI is changing how we run the business internally. We are seeing increased operating leverage across the organization while continuing to invest deliberately in R&D, vertical solutions and customer-facing functions.

Turning to the quarter. Unless otherwise indicated, I will be discussing results on a non-GAAP basis, and all growth rates are year-over-year. I also want to note that since we price and sell in local currency, fluctuations in FX rates impact results. Since the time of our last earnings call through the end of the first quarter, rates remained largely stable and resulted in an incremental tailwind to our first quarter ARR and revenue results of less than $1 million. First quarter revenue grew to $418 million, an increase of 17%. Normalizing for the year-over-year FX tailwind of approximately $7 million, revenue grew 15%.

ARR totaled $1.901 billion, an increase of 11%. This included a $9 million year-over-year FX tailwind. Net new ARR was $49 million. Normalized for foreign exchange and the impact of M&A, net new ARR improved on a year-over-year basis. Our dollar-based gross retention -- gross retention rates remain best-in-class 97% and our dollar-based net retention rate was 109%, underscoring the durability of our customer base as they embrace our genetic automation solutions. Adjusting for FX, dollar-based net retention rate was 108%, demonstrating stabilization across our business.

We ended the quarter with approximately 10,550 customers. Attrition continues to be concentrated amongst our smallest customers, while customers generating more than $30,000 in ARR grew 7% year-over-year. That dynamic is also reflected in our cohort performance. customers with $100,000 or more in ARR increased 11% to 2,624 and customers with $1 million or more in ARR, increased 18% to 374. Our customer strategy has continued to focus on deepening our presence within the world's most complex enterprises, where we see the greatest opportunity for long-term expansion.

Consistent with that strategy, we continue to add new enterprise customers with significant long-term expansion potential, including new logos like Candela Medical, Tire Rack, Shoprite Holdings and a global semiconductor company. who is replacing a legacy RPA vendor with UiPath as their strategic automation plan. Our cross-system integration and end-to-end process orchestration capabilities, given them a scalable foundation they need to migrate their existing automation program beyond task-based automation into broader agentive workflows.

Remaining performance obligations increased to $1.413 billion, up 15%. Normalizing for the FX headwind, which was approximately $9 million, RPO grew 16%. Current RPO increased to $908 million, up 17%. Turning to expenses. We delivered first quarter overall gross margin of 83% and software gross margin was 90%. First quarter operating expenses were $256 million. For the first time in company history, we delivered a GAAP profitable first quarter with GAAP operating income of $28 million, up from the prior year GAAP operating loss of $16 million.

GAAP operating income included $53 million of stock-based compensation expense. First quarter non-GAAP operating income was $92 million, representing a 22% margin, up over 250 basis points year-over-year and driven by our continued focus on operational efficiency. First quarter non-GAAP adjusted free cash flow was $130 million. We ended the quarter with a healthy balance sheet of $1.4 billion in cash, cash equivalents and marketable securities and no debt. During the first quarter, we repurchased 20 million shares at an average price of $11.47. Since April 30, under our 10b5-1 plan, we have repurchased an additional 2 million shares at an average price of $9.63 through May 27, 2026.

Now turning to guidance. We are pleased with the team's execution and what continues to be a variable macroeconomic environment. We continue to maintain a prudent outlook and guide to what we see in front of us. Since we provided guidance on our last call, the euro has remained largely state while other currencies such as INR and Romanian Lane have experienced volatility. As a result, for the second quarter and full year we expect a nominal incremental FX headwind to ARR and revenue.

Despite the incremental FX headwind, we are raising guidance for the progress we've made on our operating priorities. Turning to the specifics of our guide. For the second fiscal quarter 2027, we expect revenue in the range of $395 million to $400 million. ARR in the range of $1.929 billion to $1.934 million, non-GAAP operating income of approximately $75 million, and we expect second quarter basic share count to be approximately 518 million shares. For the fiscal full year 2027, we expect revenue in the range of $1.776 billion to $1.71 billion.

ARR in the range of $2.058 billion to $2.063 billion, non-GAAP operating income of approximately $430 million. And finally, we continue to expect fiscal year 2027 non-GAAP adjusted free cash flow of approximately $425 million and non-GAAP gross margin of approximately 84%.

Thank you for joining us today, and we look forward to speaking to with many of you during the quarter. With that, I will now turn the call over to the operator.

Operator, please vote for questions.

[Operator Instructions] Our first question will come from Bryan Bergin with TD Cowen.

Asim, maybe just to start on the overall demand environment. any interesting changes in the underlying demand trends and pipeline conversion, anything as it relates to deal timing, sales cycles, things like that, just as this conflict has been extended.

No. We actually feel like the environment has stayed relatively stable versus what we saw in the first quarter. Sorry, when we got into the first quarter earlier this year, Brian, I think we actually feel very positive about the momentum in the business, the health of our pipeline and the conversion rates and the predictability.

The customer conversations are going really well. A lot of the pilots are beginning to now starting to convert, which we feel really positive about. So overall, we're actually very positive overall on our pipeline and the environment remains variable as it has been, it feels like a new normal is the way we think.

Okay. And then on AI product AR levels, any sizing you can update us there? And how the pricing conversation across those solutions is evolving?

Yes. We'll disclose the product IRR periodically here. We feel really good about the momentum. I think we pointed to it in terms of 16 of the top 20 deals for the quarter involved AI. I think Gentex and our AI products in general have really good, strong momentum. And our vertical solutions are also starting to really get traction both from customer interest and pipeline, particularly in health care and financial services.

And then lastly, I think test, which is our Augentic testing solutions that has really good traction as well. We look forward to update the numbers here in the coming periods. But right now, we feel really good momentum. And I think the deal traction kind of speaks to the overall trajectory for the [indiscernible].

Your next question will come from Scott Berg with Needham.

Hi, everyone. Nice quarter. Daniel, you spoke extensively about orchestration, and it's a key topic that comes up in our work on the space consistently over the last probably year -- when you think about Mitro and the deals that you have out there, is there any reason why Maestro is a part of basically every deal that has AI? Or is there some combination that would suggest that, that's not going to be a part of every deal going forward.

I don't think my store can be part of every deal. The way we are looking at our business, it's -- we have an entire platform that can address the whole spectrum of past can process orchestration. Maestro is a solution that comes into play when customers are doing process orchestration and automation and end-to-end process orchestration and automation. But we have customers out there that are happy to start with the task automation product.

And thus automation can also be deterministic and cabinet. I would say that RPA and API automation plays into deterministic task automation, while we have agents that can be applied to task level. Maestro comes into place when you need more complex orchestration of work that involves humans, task automations, enterprise workflows systems and agents.

So it's naturally more for our more involved customers. Maestro helps us lending bigger deals. -- makes our installed base stickier to the customers, but I cannot say it can be deployed in every single year.

Got it. Helpful. And then Ashima follow-up to the last questions that were out there. I think what we're all trying to understand is the impact of, obviously, some of your AI modules on the business and the bookings and what the general trajectory is. I understand that you don't want to necessarily report that AI metric every quarter.

But if I ask a question a slightly different way is if I think about those 16 deals in the top 20 that had an AI component of them., How significant are those transactions is coming from some of the AI functionality. I think we're all trying to understand is it still traditional RPA heavy in those transactions or if we're seeing a bigger impact from some of the AI function...

No, we're seeing a bigger impact I think the way I look at it is I kind of would divide it into 3 areas, like our top customers and our CoCdeals, the majority of our transactions have a significant AI, if not a majority, AI component. Scott that's driving it. They're not piecemeal where it's kind of like 1 or 2 SKUs that get moved in or small quantities. They are materially what we are selling, right, to our customers. I think there is a mid-tier of customers where you see actually a continued demand in traditional RPA and deterministic automation.

And those are companies that are not -- that either are -- have embraced in genic and AI in a major way, and they are actually pulling forward more deterministic automations as they weigh both the cost and the trust and governance. -- that agents versus deterministic automations give you. And then really, the kind of some of the drag that we talked about is really from the low end of the market, smaller customers and personal productivity. That's kind of the way I would divide up the quarter.

So we're actually really pleased with the pull that we're getting on the Agentix side and its contribution to our growth.

Your next question will come from Sanjit Singh with Morgan Stanley.

This is Abhishek from early on for Santen. -- just to hear a little more on the beat kind of dig into -- given Q1 revenue upside was strong, but the beat was largely driven by license revenue and the ARR was relatively in line. So can you kind of help us understand the quality of that revenue be -- is there anything unusual in license timing or customer behavior that we should be aware of? And then how should we think about the relationship between license performance and ARR trajectory for the rest of the year?

Yes. I mean, I would say 2 things. One is we feel really good about the quality of that revenue, both in terms of the products as well as the deal quality and structures. I would say it's -- our quarters have been very clean. And we feel very good about the overall deal quality and construction. Remember, revenue is a quarterly performance metric when you're looking at the growth rates, and we are on ASC 606. -- versus ARR, which is a 12-month metric, right?

So if I break down the question, you look at revenue growth at 17%. When you look at a trailing 12-month period, the revenue growth rate -- so actually, which makes me feel very good about 15% growth on a trailing 12-month basis. And it's relatively in line with the ARR growth -- in terms of ARR versus revenue need, it's really just the mix of deals with 606 timing. And the license revenue being a factor in that is a side actually of really good quality revenue overall.

And then as a follow-up, anything you can share in terms of the mix between consumption-based revenue and proceed?

We don't -- consumption-based revenue is a very small part of what we do. We still have -- the subscription really dominates our pricing model. and per seat pricing as well, that is not the majority of what we do. We are really selling executions as well as kind of our typical server-based pricing that we have for unintended robots in particular. I just really emphasize again, personal productivity is a very small part of our portfolio, simple task-based automation. So what we sell is the larger complex use cases now -- and that really mix us higher towards both server-based and subscription-based pricing. .

Your next question will come from Samik Merchant with RBC.

Guys, this is Sonic Mojin on format Hedberg RBC -- could you talk a little bit about the broader competitive environment for orchestration and any changes or trends you're seeing -- and there's also been a lot of developments around frontier model capabilities. Could you talk to how you see these developments impacting the broader competitive landscape and the company specifically?.

Yes, sure. I would like to start by saying that we have a really unique platform in the market. So -- and it's based on 3 major viewers. We have a very modern process orchestration technology that is built on a very innovative workflow engine capabilities. We have proven a 10 years deployment of scales of automations in a secure and governed environment with some of the largest companies in the world. And we have a unique ability to connect to both modern API-based systems and legacy systems. This 3 pillars make our platform quite unique in the market. In terms of the new development that we have seen I think we all recognized the huge impact of the coding agents of the entire ecosystem.

And I want to point to you to an interesting phenomenon that it's -- it's something that we spot with our customers and within our own IPP operations. It's becoming increasingly easier to build deterministic automation. You are using coding agents to build deterministic automations and deploy very much scale. It's becoming really easier to address the long tail of opportunities of work. And it was not economically feasible before coding agents to get to this level of automation. And building automation, it's really creating the substrate for deploying genetic AI later on. I would point to why coding agents are so successful nowadays because they really combined model, the strength of the models with the strength of deterministic automations.

[indiscernible], it's so good because there is this deterministic harness around it. So cloud generates cost but then it uses a compiler, which is a deterministic piece of technology to compile the core, and then it's using testing, which are another deterministic piece to validate the code that is generated. So I think it's becoming more clear to everyone that the combination of deterministic automations and models are what makes the real deployments in production. And I would say that in this regard, we do have tremendous advantage -- our platform is already enabled for coding agents, and we showed that our decon in India, we show that we can reduce significantly the implementation times. Think for a second weeks, 2 hours, that really means a lot when you go and deploy automation to the long tale of possible work.

Thanks. Appreciate the color there. And as a quick follow-up, so you've talked a lot about sort of profitability. And last quarter, you also updated new long-term non-GAAP operating margin target to 30%. And keeping in mind the fact that we remain a priority for the company, what are some keys to margin expansion in fiscal year 2017? And is there any seasonality you would point out on that?

Look, I think from a cost seasonality, nothing except for -- obviously, there are later parts of the year, we have sales compensation. There's just normal SaaS seasonality from an expense standpoint. Otherwise, I think we're pretty -- there is no real seasonality to mention. From my standpoint, I think we're looking at as growth is our first priority. So we are investing in FTEs. We are investing in test. We're investing in vertical solutions. We are investing in coding agents as evidenced by the speed of the launch by which we're moving through things. And so from our standpoint, that investment is our first priority.

At the same time, we updated our long-term models because we are able to find increasing levels of efficiency both through continued discipline and scrutiny and then also from implementing both our platform as well as broader AI tools within the company. And so I would say we're a best first mindset. -- and a waste nothing mindset. And that combination, I think, gives us the ability to both grow and drive the strategic initiatives while expanding operating margins.

Your next question will come from Pat McElwee with William Blair.

Daniels. My first question, I thought the AI summit you put on earlier this year was very helpful in envisioning how customers can evolve from your traditional RPA workflows towards more agentic-enabled workflows. And specifically how they can choose their own autonomy level and then kind of use a feedback loop to evolve the level of automation in that process. over time.

So I know it's early on, but for your existing customers, where are they in that autonomy, evolution right now are a lot of them content with the value they're getting from current RPA workflows and leveraging AI within newer workflows? Or are they really racing towards these agentic solutions to maximize the ROI they're getting from the platform, both existing and new workflows alike.

Yes. I would like to point out that despite the technology being very new, it is hailed by our customers with a lot of enthusiasm. Even when we were in like close review, we got a lot of requirements from the customers. They -- some of the customers even went to find only some of the skills that we publish and use them with the coding agents. And Also, I would like to point out to the fact that basically holding agents saw 2 of the biggest hurdles in deployment of automation. .

Number one was always the implementation leading time to when -- until an enterprise would get value from automation. So -- that's been already proven internally by our own forward-deployed engineers and externally by a few advanced customers that can be shrinked in many cases, from weeks to hours, which is very significant. The second way that coding agents unlocks a bottleneck of automation is in maintenance. One of the apparent flows of automations was always the fact that they are project and they break.

If there is an upstream modification in an enterprise system that automations are not if they might break and that will require human interaction and many days of reviewing and understanding. Now we offer both a healing agent and the diagnosed agent. So the heating agent can do a lot of the work during run time during execution.

And in many cases, the hearing region can fix the execution in itself and the processes are unaffected. When there is an exception, we help tremendously developers with these diagnostic agents to gather all the context around automation, and they can publish a fix in much faster than before. So yes, I would conclude that for us, this is a really big unlock. And we see the potential for a huge acceleration of customer adoption.

Right. Okay. And to kind of continue on it, it sounds like AI agents are largely extending, not replacing deterministic automation within your platform. But as we think about that, I wanted to ask, is there any sort of dynamic where you're seeing customers leverage Gentek AI to somewhat cannibalize some of the traditional bot monetization? Or is it largely building incremental automation and therefore, incremental monetization on top of those workflows?

Yes. I would like to say that perhaps this is 1 of the biggest confusion that AI brings into the table. The idea that nondeterministic probabilistic technology can replace deterministic automation. This is not true. It's not true from the capability perspective, and it's not true from a economical standpoint. And let me elaborate a bit on both. A probabilistic technology is not architectural meant to follow a dozen of steps and sometimes hundreds of steps in the same order in the same sequence. Every step will have a probability.

When you multiply these probabilities you will end up with something that is not reliable end of the day. And there are many regulated industries that cannot tolerate anything that is not 100% reliable. They will prefer an automation to pay as an exception rather than produce an unexpected result. So what that [indiscernible] bought cannot be replaced by nondeterministic agents. Again, this is the architecture that is proven over and over again by all the AI agents that allow them, the most sophisticated agents like lot codes were open AI codex are built on the foundation of deterministic tools that they quote. So it's a hardness around the model and deterministic tools. This is exactly how they work.

This is exactly what we are proposing to our customers, yes, reduce your investment in your existing deterministic automation and surround it with process orchestration, which is also deterministic that are orchestrate models, agents in the context of determinism. That's really the only way to deploy effectively into an enterprise context. And now to the second point about the economical aspect, even if in certain cases, a major can replicate some steps that are [indiscernible]. Why would you do something that is costly and it's going to consume tokens at every step in the process rather than generate a script that works. It costs nothing in order to run.

So to my previous answer, this is the best combination between AI and deterministic. AI creates automation, sometimes maybe even on the slide. I will run those automations. It's very cheap to run, very that we stick reliable or deal and only when this scripts break, you can invoke again a high to fix the screens, but that's basically the right model to run agentiKI and automation into enterprise context.

Your next question will come from Raimo Lenschow with Barclays.

Perfect. Daniel, could you stay on that subject because that's obviously where a lot of the investor questions are coming around. So how do you -- how does the world work then going forward? If you do the deterministic part, and you have all the experience in the words, so you should do that, who is doing then the probabilistic part. What are you seeing there in terms of where customers thinking and how they think about you in that context? And then I had 1 follow-up.

Well, I think the answer varies in we are model agnostic in terms of how we see the old -- we provide deterministic orchestration and we can infuse that monistic at orchestration at any steps with a GenDKI.ThataKI used behind the same frontier blood models can use open weight models. We have to bring your own model policy. So we will accommodate every spectrum of requirements from our customers. But again, I think what's important to note that even on the frontier lab model, the offering, it's a combination between deterministic and the model itself, which is purely cognitive.

We extend in a way that model into the enterprise work itself. And when you go and I think very important distinction to understand the enterprise work is to think of who initiates an agent or process automation. It's a big difference if it's initiated by a person and the agent work on a personal desktop versus an automation is triggered by an event or buy an enterprise workforce where you will need to have a different degree of auditability and reliability. And again, this is where we really shine. We have this 10 years of experience in running a large-scale unattended automation that work on event figures. And we are involving agent KI and models into these workflows that can run unattended.

Yes. Okay. Makes sense. That's very clear. And then in the 1 other question I get from investors a lot at the moment is you're doing really well on the revenue side. ARR is very steady. But at some point, they kind of need to kind of start lining up. So revenue at the moment keeps growing faster than ARR. How do we need to think about that dynamic? Because in theory, you would think that they should line up or we think..

Yes. I think, Rama, the first piece is, again, like when you look at it on a trailing 12-month basis, the revenue growth rate is 15% versus the ARR growth rate of 12% that you see. The second piece is within the revenue growth rate, there's obviously the license revenue growth rate and then there's services, et cetera. You can see we actually had good services revenue as FTEs, et cetera, are in demand from our customers. So that's a second piece. -- that is there.

Over time, this has moved in different directions. There's been times where with 606, revenue has trailed as certain duration and mix has moved the growth rate and where it exceeded -- but when you look at it like on an overall average over a longer period of time, it's together.

I don't really see any major disconnect at this moment that is driven by a business-specific area. It's really just a mix of 606 impacts on the business. And again, I would emphasize to look at it on a trailing 12-month basis, versus looking at it where it is just in a particular quarter because ARR is obviously an annual metric.

Your next question will come from Michael Turn with Wells Fargo.

Reappreciate you taking the questions I'll just ask 2 upfront, and you can take them on to sequence you like. I guess the first is just in terms of public sector. as we roll into midyear, maybe just remind us how you're thinking about public sector this year. Any updates in terms of progress or deal progression from that vertical specifically and maybe option just on the net retention rate, just what you're seeing currently in the uptick there and how to think about the trend line obviously, without guidance, but just thinking through the drivers there.

Yes, I can answer both questions. I'll start with the net alternate. I'm actually super excited with the Nutella retention rate and the progress we've made on it. As you can see, we have a 2-point increase quarter-over-quarter. That's 1 of the first times we've had an increase as we've stabilized net new ARR and beginning to point the trajectory up towards that reacceleration Mark. One, when you normalize for foreign exchange and the impact of M&A, -- that is 1 point, but it's still a reacceleration of net dollar retention. I we're actually very encouraged by. And as I said, as we start to stabilize net new ARR, the next step is reacceleration.

So we're moving into that territory. And I think that's really great progress by the teams and speaks to the strategy and the operational execution that we -- that you see. In terms of Poly, I actually was at the public sector, a fusion event that we had -- the energy was very strong. I think public sector in terms of disruption of budgets, et cetera, we feel pretty -- we feel like there is good stability.

Obviously, as funding moves with different defense initiatives and awards, et cetera, that are there. We stay on top of it. But within many agencies, we actually have a very good presence, strong relationships with really good use cases, whether that is audit compliance within the government, which we have a very strong set of solutions and partners that we're working with or other transactional areas we actually feel like our relationships are very good. In terms of what's in for -- as we talk about , we continue to guide what's in front of us there, which is we're pretty measured and prudent. We know what projects are generally funded then we're looking to execute against that.

Your next question will come from Radi Sultan with UBS.

First for Daniel, just on the UI Path for coating agents, you mentioned this will be targeted at the full software development life cycle. But I guess -- are there like 1 or 2 areas where you see the biggest pain points where you can add the most value. And then you also mentioned this could strengthen retention. Maybe just how you imagine monetizing or bundling these agents into the broader suite.?

So we plan to bring agents across the entire development life cycle. We are starting with an agent that helps us planning for an automation. So you can have a business analyst that came with the help of the agent interview different subject matter experts, gather all the information, create a process documentation document -- and then we will have like a solution architect agent that will take this design document and convert it into code. And we will have different agents for different types of course.

We have an agent that can create enterprise user interface. We have an agent that will create RPA, another agent that can create API of course, an agent that will create process orchestration based on milestone -- this can be deployed and tested, again, it's fully agented. Once they are in production, we have agents that monitor the entire execution and can fix proactively the errors that are coming. And once there is an exception, we have also agents that help our developers to diagnose faster, the exceptional fixed them faster.

So the entire life cycle, there is no single point in the life cycle that is not touched by pages. In fact, we believe that the entire offering surface of our platform is basically a genetic first. Humans, we think, are mostly going to do validation. They will inject goals to the agents, and they will do the validation and supervision of the work. But most of the work itself is going to be created by agents.

Got it. Got it. Maybe just 1 follow-up for Asim. Last quarter, we talked about core RPA still growing and becoming increasingly strategic to the AI product offering. Can you just talk through how you think about how pricing should evolve for the RPA and terministic automation side of the business, given that it's becoming increasingly more strategic to customer AI initiatives to kind of capture that incremental value?

Yes. Look, I think that there's a lot of discussions around outcome-based pricing that are real and active more than they ever have been before. So I think like that is 1 tier of pricing to our top customers. that I think it's a real evolution. We see real lined path. And we see people, especially with their fears of -- about getting ROI with AI. -- really looking for that.

The second piece I would say is I think where we're looking through is we also see like use case or process-based pricing. -- where people are looking for restricted use cases so they can solve problems and have -- be able to use different parts of the platform that enable them to do so. Those are 2 evolutions that are there. in terms of where we are with the Determine side and overall.

This now concludes the Q&A session. I'd like to turn the call back over to management for closing remarks.

Thank you so much for the questions. And as usual, we would like to speak directly who many of you over the next few days. Thank you so much.

Hi, everyone. Thanks for joining us today. I'm Allise Furlani UiPath's Investor Relations team, and I'd like to welcome you to our virtual fireside chat and product strategy overview. We'll begin with the fireside chat featuring Daniel Dines, UiPath's and Chief Executive Officer; and Raghu Malpani, Chief Product and Technology Officer. We'll then take a deeper look at our product strategy and road map, followed by a customer example to bring these concepts to life. We'll conclude with time for questions. [Operator Instructions]. Before we begin, I'll cover a few housekeeping items. Today's event is being recorded and will be posted to our Investor Relations website following the session.

I would also like to point you to our safe harbor statement and remind you that today's discussion may contain forward-looking statements. Actual results may differ materially from these statements as a result of various factors, including those found in our SEC filings. We may disclose information related to development and plans for future products, features or enhancements, which are subject to change at our discretion without notice.

All statements are made only as of today, and UiPath undertakes no obligation to update any forward-looking statements and makes no assurances and assumes no responsibility to introduce future products, features or enhancements described today. Additionally, we would like to note that this is a product webinar and we will not be taking any financial questions.

With that, I would like to hand it over to Daniel to begin. Daniel?

Thank you so much, Allise, and hello, everyone. Thank you for joining us. Today, I have the pleasure to introduce Raghu to you. I heard many people in my life. And -- but in very few cases, I had really this type of positive vibe that I had when I met Raghu first time. He strike me as a clear thinker, no nonsense, no politics type of guy that we really needed to run our engineering organization, and he didn't disappoint. Raghu was basically the artisan behind our big push into Maestro into adopting the new modern workflow engine that is changing the entire company.

So, Raghu, what brought you to UiPath, basically?

Yes, Daniel, thanks for the kind words. I appreciate it, man. Look, as many know here, I worked at Microsoft and Facebook prior to coming here, worked at Microsoft a few times, left it a couple of times. Worked on Azure -- worked on what became Azure in my first rodeo there. And then the last time I left, I left the office organization to join UiPath, to join you and the team here about 2 years ago.

After spending close to 20 years in the industry, I wanted a place which met some expectations for what is it that I work on? Who is it that I work with? And how is it that we work together? So Dan, you and I had many conversations before I joined. In fact, we met, I think, 2 or 3 times in person. And then it was -- it is clear that the transition from RPA to Agentic automation to business orchestration was the need of the hour. The ACT I and ACT II transition that you've spoken about for a couple of years now made this to me a compelling -- a resoundingly compelling opportunity, actually. And where my passion and my interest aligned with helping drive a significant transformation, working with people I enjoy working with.

But more importantly, I think -- and we don't talk about it enough in the industry is what also drove me here is who I work with and how we work together. It was clear that Daniel's way of leading and the culture inside of UiPath best matched my ideal workplace. It's a uniquely customer-centered organization where there is humility all around with everyone you work with regardless of levels and titles. Being plainspoken and direct, being humble and decisive, being strategic and hands-on at all levels, where everyone generally steers in the same directions was an important part of how I wanted to spend the next decade of my career. So working in a [indiscernible] environment, making meaningful impact. And I found this company, UiPath to meet the cultural needs for me to thrive doing the kind of work that I was passionate about.

And I'm so glad to be here, Daniel, actually, it meets all of my expectations. As we all know, the world of software is changing so rapidly. The winners and losers are to be decided, to be quite honest. But the technological moat that we have and we are building, the customer base we have and we are growing. And most importantly, the culture of fast decision-making and the velocity and the customer centricity, I think we have the ingredients to build a great company and strengthen the great company that we already have. And so I'm excited to be here, and it's been a fun couple of years and I'm looking forward to many more, Daniel.

That's great having you here. And you passed the best test of our engineering leads in Romania, that was part of the hiring process. So I was impressed that you are the only one or maybe two people that got a good review from that site.

Yes, yes. It was both hard and easy to pass the test. It was hard the first 5 minutes, but I knew who I was speaking with and the connection was deep and immediate, Daniel. So I think I understood in those first few minutes talking to some of those folks, what was expected and the cultural alignment was clear, clear from the get-go. So I'm...

And you were talking about the [ code ] in the interviews.

Yes.

Myself, I have asked you some kind of hypothetical coding stuff, so good. So I think for -- it was really a good lesson how to get someone culturally fit.

Yes, it's super important, Daniel. And I think the leadership team that we have at UiPath now between you and your direct reports. I feel like we are -- we have that team where we can speak freely, we can spar openly, challenge each other and really push the boundaries for what is possible. So I feel like we're in a good spot for all the reasons that we just discussed.

Daniel, I want to ask you a couple of questions that I'm sure is top of mind of many investors that are here. There's is a narrative that AI could simplify or even eliminate large parts of the software stack. In that world, what moats do you think we as UiPath have? And how does our role become you think, even more important? Do you want to take a stab at that and then I can add my 2 sides?

Yes. Yes, of course. Look, as we all know, this is not a new narrative for us for UiPath. So I think we've been on this AI kill list, the first, I think, around 2023. And then we've been on and off that lists sometimes we've been on the benefactors of AI, sometimes on the kill list. So I would like to debunk maybe a few things about AI versus RPA and versus UiPath.

So look, I think it's important to distinguish between using AI during implementation time, during design time, during process requirements definition and using AI during execution. Because I think first of, the main important question is AI capable of executing a task as RPA is doing. Well, the answer -- it's very nuanced. In some cases, it's capable in many cases to run a complex multi-step task, growing multiple application completely autonomous, it is not today. So the most interesting incarnation of that we are seeing have seen, like Claude Code, so they are meant for doing ad hoc tasks in the presence of the humans. So humans is ultimately what decides if the payment transaction is processed or not.

Our business is really about running complex processes, workflows in an unattended autonomous fashion. And I think this is also even philosophically, it's not the territory of AI. AI creates a code that is running on an infrastructure, it's running on the framework. But it's not meant to replace that code even if it can. In a very simple example, AI can multiply 2 numbers right now. Look, sometimes, there can be errors. If I really put it to test and give it very big numbers, AI might not succeed. And anyway, if I have to multiply 2 numbers in a loop a million times, AI will struggle, and it's not going to have 100% accuracy on that one. Nobody ever is thinking that AI should multiply 2 numbers, everybody really understands that AI will call a tool that will multiply these 2 numbers. AI is not enough to understand, this is a request to multiply 2 numbers. Therefore, I'm going to call a tool. We think it weights hard to translate it into automation. Obviously, AI can understand when it's the right context to run an automation and it's going to run that automation.

How do you create that automation? This is where we shine. And we offer amazing platform that makes it very compelling for most of the enterprise to build their automations to run on our cloud. And also as part of this, it's not as simple as you create one automation, you run it and you are done. This is -- if you look at the landscape of enterprise processes, it cannot -- it's so much complexity there. When we speak of a process like procure to pay or order to cash, we might have hundreds of sub workflows involved that have to be clearly orchestrated by rules, by policies, by human judgment. And so you'll have always multiple actors. It might be a dozen of people involved in a process.

It's not -- people are not going to let just the black box AI do my order to cash and they people and AI have to reason over a system. So right now, what I'm the most excited is actually the emergence of the coding agents. Because this is basically -- it's the best of both worlds. Coding agents will help our customers and our partners to build automation. And to build automation at a larger scale that we've seen before. This is basically our biggest road map change that we had in the last few months. So we pivoted the entire company into enable our platform to be used primarily. We even see the primary person to use our platform is going to be the coding agents.

And we will support where coding agents agnostic. Of course, we work with Claude Code, we work with Codex, and we will work with all the best coding agents out there. But our ambition is to have our platform enabled since the inception of an idea of automation to creating a process specification, interviewing multiple stakeholders, understanding the process, the nitty-gritty of the process, creating the solution architecture, creating all the artifacts, including RPA, including Maestro, including document understanding to debugging, testing, deploying in production, monitoring in production, fixing all the exceptions that happens in production.

So AI will be the main factor in interacting. So it's going to be a very heavy use of AI, but who runs the pieces, the artifacts that run on our platform are good, are very reliable, work a million time in the same time, you can reason the same input is going to produce the same out. It's governed, it's secure, it's auditable. So in a way for us, I now I see that really coding agents, it's an amazing accelerator.

Yes. Thanks, Daniel. So how about -- it might be a good segue for us to talk a little bit about our product strategy, Daniel. Talk to our investors on how -- like you mentioned, how we're going up that value chain from tasks to more complex processes and then how coding agents make -- is a massive force multiplier for a platform. So let's just jump right into the product road map presentation. I'm going to share my slides.

So let's talk a little bit about what Daniel spoke at a very level for the vision of where we want to take our products over the coming months. We'll talk a little bit about our product strategy and how we believe we capitalize on the opportunities that are ahead of us in the coming year.

All right. So let's first maybe look at the reality in the enterprise today. There was a survey done with a few CIOs was last month, where you'll see on the left that integrations is the #1 barrier for transformation, agentic transformation and automation transformation. Nearly 50% of the CIOs say that connecting AI agents to existing systems like their databases, their CRMs is in our challenge, and it's compounded by all kinds of data quality issues. And on the right, you see the emergence of -- this surprised me, too, the emergence of shadow AI where about 20-odd percent of deployments are already unauthorized. Now this makes it clear that enterprises need a unified platform that solves both sides, the deep integration into all types of systems and governance built in.

Now of course, UiPath solves for it, but it also clearly gives us signal that the best way to solve the problem is to really go up that value chain to take away this complexity from the CIO, the COE and even the business users and provide them more turnkey and end-to-end solutions.

Now let's take that enterprise reality and transition a bit to talking about how we are going to take our platforms and our products forward. We want to introduce to you our agentic business orchestrator platform. We'll read this from bottom to the top, you see -- we started where UiPath found its first product market fit and defined category leadership, as you know, at the task layer with UI automation and pretty all-encompassing enterprise connectivity. This is the foundation that let us reach into any application and any interface. It's our biggest and deepest moat, as you all know. And it cannot be overstated because the most complex enterprise processes have a mix of legacy and modern applications, and it's critical to have robust RPA support and API support to meet that diversity that we need.

Now then, as you may know, in the last 12 to 18 months, we've moved up that value chain to also include process orchestration in our product mix with Maestro coordinating robots, humans and AI agents, to orchestrate the most complex business processes end-to-end. We've built some of the deepest technological moat here with support for long-running workflow, that run for hours, days, weeks, even months spanning multiple agents, multiple systems and humans, work that gets interrupted, work that can fail, work that needs to recover and still complete with full visibility into where everything stands at our throughput, auditability and security, some things that enterprise customers really care about.

And then further up that chain is what we call agentic case management, where we embrace the dynamicism that exists in business processes and the countless variations that they take. These processes are chaotic to say the very least. And our newest innovation, what we call our case manager agent can autonomously triage and resolve these complex and dynamic processes. This is our newest offering, launching in May, and I'll talk a little bit about it in an upcoming slide. And then we are now at the top of the value chain here on vertical solutions, where we offer industry-specific offerings that deliver and outcomes to customers that implement the highest value business processes end-to-end. And also, we'll cover this in a little bit. Now that's really the value chain story. We're going from tasks to also include processes with our orchestration layer to cases and vertical solutions. All these outcomes delivered on our secure and governed platform.

Now I'm going to spend a little bit more time talking about the modern agent native stack. As Daniel described a bit earlier, and you've probably heard that the coding agents are taking the software world by storm, and they are totally revolutionizing. Our software is built and managed. Now at UiPath, we are going all in and making our platform fully accessible by coding agents. This includes the full life cycle of automations. From building, operating, managing and obviously, of course, governing automations as well. So on this slide, on the left, you see our platform stack, which I just talked to you about.

On the right is the key unlock. This is coding agents natively embedded in the platform. They accelerate every phase of the automation life cycle. They will trust this context, reading process documents and really capturing the requirements for a real business workflow. They author and deploy, meaning they go from natural language to production-ready agentic workflows with guardrails. And then they diagnose and repair proactively analyzing logs, errors, proposing fixes even and redeploying them in a closed loop. And then critically, they support governance and operations, managing machines and making sure that SLA, business SLAs are met and adhere to. The key here is that coating agents compress the time to value dramatically. Every developer, every operator on the platform becomes more productive. What does that mean? That means that more automations get built secured and governed. It means that we really can expand the value inside our existing accounts and accelerate new customer onboarding as we add new accounts to our list.

Now there is a segue that I'll present here, which is our developer base expands also. Today, as you know, we target what we call automation developers. But coding agents allow us to target pro developers in the enterprise as well as those that are less technically proficient. Now does this mean that our bets and investments on low code experiences go away. But we don't think so. It actually becomes more important because it gives people, especially those that have less technical proficiency, the confidence to visually verify and inspect that their natural language intent matches the actual automations were done. So we believe that the low code experience becomes even more critical, at least in the short, medium term in terms of how people express and manage their intentions.

All right. So we've introduced the business orchestration platform. Now I'll spend a few minutes taking a deeper look at the orchestration layer, the case layer and the vertical solutions player. Now this slide describes the very core foundation of our orchestration platform, the orchestration layer. On the left, you see a real orchestration flow, this isn't, as you can see, a linear automation. Even though it's simplified to fit in this slide, you see that it's a reasonably complicated process with branching logic with human checkpoints, multiple agents and handoffs, all coordinated by Maestro. On the right, I'll talk to you a little bit about what are the sets of salient capabilities we've built in here that makes this enterprise grade.

First and probably most importantly, it's an integrated and unified platform where you can model your most complex end-to-end processes, but also implement every single constituent offer. To be clear, Maestro is endpoint agnostic. What we mean by that is you can bring your own systems of record, you can bring your own agents built outside of UiPath, it doesn't matter. The problem we solve this orchestrating these complex processes, planning systems, agents, build on multiple platforms. Somewhat still needs to move the process along and execute it optimally. Someone still needs to provide you the visibility on how the process is doing. And that's the core of Maestro. It's like that control tower that gives you that visibility, that drives the process along to completion.

Second, this is our technical moat, which is the durable execution. When an agent fails, when a system goes down, especially for these complex processes and they happen all the time. The orchestration layer needs to pick up exactly where it left off to ensure that we provide mission-critical reliability built on this engine we call the event source engine. That's really at the gut of this new wave of products that UiPath is building, Maestro, case management and of course, all of the vertical solutions as well. And then, obviously, business users stay in control, humans are in the loop at the right moments for triaging escalations, approvals and so on. Every step of Maestro is audited and governed end-to-end, so that -- as I mentioned earlier, there's concerns from the CIO is about governance, and this directly addresses that governance concerns that they raised.

And last, but critically probably equally important to all the rest is Maestro has or will soon have native integration for coding agents. So these orchestrations don't take months to build. Developers can use coding agents to also test and deploy them quickly closing the loop on time-to-value story that we just covered. All right. Now let's get real about what a business process actually looks like. This is an actual insurance claims process and really just one variation of it. Look at the complexity. A claim comes in, an AI agent runs evaluations it branches, maybe the confidence is low, it escalates to a human, the adjuster may correct the data. It loops back and there's probably some legal and compliance coverage or reviews that need to happen in parallel. Maybe some parts of the process needs to be reevaluated and so on.

Notice the nodes here, there are AI agents participating in this process. There are some deterministic automations, APIs and RPA automations of the mix. The nodes that are dark red, the red-shaped, the red-colored nodes are the humans in the loop. A single process leads to all these 3 components, agent, humans, APIs and robots continuously. It's not about really automating a single step. It's really about orchestrating all of these constituents across all of the variations and exceptions and making sure that the work actually completes. This is why naive or simpler workflow tools breakdown, they can handle the happy backlog. But the exceptions, the handoffs between AI and humans, the parallel branches, that's where you need a true orchestration platform, and that's exactly what we've built.

Now what you saw in the previous slide was the reality, right, the messy and tangle process. The goal really isn't to create that perfectly linear flow. We know it does not work. The goal is to orchestrate everything that needs to happen. When and as it needs to happen with Maestro's agentic case management capability. At the core of this is our newest technological moat, what we call the case manager agent. This is a foundational investment that makes our push into these complex processes possible. What does it do? It maintains and manages state in context and progression of work across all of the stages that you've seen here.

Think of it like a brain that knows where every case stands and moves it along. As you can see here, the process has broken up into these 3 stages, the intake step, where claim comes in, an agent processor, an AI agent extracts the document, the case manager agent decides the path dynamically and not using a fixed flow chart. They do with the second and third stages. This is the same complex process, same variation, but now it's orchestrated and governed and observable end-to-end because the case manager agent holds it all together built on top of our orchestration layer. And remember, even this is coding agent power, like I mentioned earlier.

Moving on to -- and this is a quick glance at how this manifests in our product, we'll see a demo of it in a minute. Now moving further along the value chain to providing solutions, as I mentioned earlier. This matters because many enterprises do not want to buy AI in the abstract, they really want to buy outcomes. They want fast cycle times. They want measurable ROI. Our solutions are explicitly designed to make value visible in days, not weeks and quarters. And this is not a separate strategy from the platform. It is the platform. Every vertical solution is powered by the same underlying platform that I just shared with you, the same governance layers, the same AI trust layers, the same data and integration layers run underneath. It's key to note that we are not pivoting to solutions. We are expanding our addressable market upwards from selling infrastructure to IT, but also selling outcomes to businesses now.

It'd be fair free to ask what makes us uniquely biased to succeed with this? And our answer is a simple. Our core thesis is actually pretty simple. AI, as you can all -- as you all know, we'll reengineered every major business process. And we believe we are uniquely positioned to lead because we combine deterministic and agentic logic in one credible platform with strong governance supporting complex and regulated areas and systems. And what we're building are solutions, not point products. The architecture, as you can see in this picture includes what businesses care about. Domain expertise is built in, specialized agents that understand bespoke industry-specific logic and bespoke business-specific logic, workflows pre-configured for use cases and ROI dashboards that speak the business users' language correctly. And we're also disciplined about how we will scale. Instead of attempting massive transformation, we start with high-value subprocesses where we can prove impact quickly, build trust and then broaden from there.

And then here, we are populating the previous picture with a few vertical solutions that we are investing in. On the left, you see by industry where you'll see our investments in financial services, health care and life sciences. These are our customers that we have -- these are industries where we have a proven strong customer base, and we understand these industries deeply. On the right, you will see the departmental level use cases, QA testing account, accounting and procurement, Test Cloud is our beachhead into the QA department already a leader in the Gartner in Forrester quadrants, and we continue to see a significant momentum there.

Now the key point is each of these solutions is built on the platform, as I mentioned earlier, this really means that every new solution we ship makes the platform stronger. And then the platform becomes stronger with every new solution. It's that compounding model that makes the solutions and the platform stronger over time. Now I'd like to make a lot of this real to you. I'd like to invite Mark Rubinstein, our Director of Product Management, who is walking through a real example of this vertical solution that he's up leading. Mark, why don't you explain to us the product that you built via demo and then explain to our folks here. So take it away.

All right. My name is Mark Rubinstein. I'm helping lead our vertical solutions team on financial services. And over the last several months, we've spent time with dozens of lenders sitting alongside loan officers, processors, underwriters, QA analysts watching how the process actually works for loan origination. And on the front end, before underwriting, loan officers and processors manually collate dozens of documents. They're hunting for gaps that could stall underwriting and they're repeatedly circling back to the borrower for more information.

And on the back end of the process after underwriting, QA analysts work through hundreds of business rule checks. They're manually leaping through dozens of documents to check and catch compliance or data entry issues before closing. And ultimately, the result is that there's no single source of truth, cycle times drag until one of their competitors end up closing faster and winning the business and cost spike every time volume surges and risk keeps accumulating with every file that relies on humans to catch it. And similar to what Raghu showed earlier, the process is nowhere near as linear as it looks, as I showed on the last slide, it's extremely dynamic. It's exception heavy and fragmented. There's loan origination in one system, core banking in another, documents in another, checklist in another, there's no orchestration layer that's connecting them, humans, these separate teams of humans are the glue that hold it all together. And that's exactly why there's errors, delays and high costs.

And this is what their day-to-day actually looks like. It's multiple systems that open simultaneously. They've got documents scattered across tabs. Data is being manually cross-reference. I can just feel their pain looking at this slide. And this is the environment that our solution must work within. And these aren't just operational headaches. They show up directly in the numbers, with 42 days on average to close a conventional mortgage, nearly $11,000 to originate a single loan and 2/3 of that cost is labor, and 47%, almost half of critical defects that are found for these loans are directly tied to manual verification and calculation. This is the cost that the process has that really hasn't fundamentally changed for a wide gamut of our customers.

So this is where our UiPath solution for loan origination comes in. It has 2 purpose-built modules. We have loans set up on the front end of the process between application and underwriting. It automatically reviews loan data and documents, identifies gaps, recommends remediation and it helps expedite borrower follow-up. And then we have the QA/QC module that sits after underwriting and after closing as well, that ensures that documents are clean. Business rules are applied consistently, escalations can be handled efficiently and the lender is audit ready, and I'll demo this module in a bit. And both are connected directly to existing loan origination systems, content management systems, core banking systems, there's no rip and replace needed, which is very important.

And together, they're designed to cut setup time in half, cut QA/QC review from hours down to minutes, so that they can lead to faster time to close, lower ever overhead loan and fewer defects. And all of this is coordinated using UiPath Maestro, automations that pulling loan data and documents from their systems of record. We have agents that extract relevant fields and execute hundreds of checks. And then there's people that can operate this solution in a single workflow, which again, I'll show in a bit. None of this was built on assumptions, by the way. It was all co-designed with a set of real customers deployed in real production environments. We started with regional banks and credit unions, some of which are shown on the slide here so that we could move fast, we could learn, we can iterate quickly. But we're seeing the same challenges at significantly larger global banks and we're working to onboard more of these customers and expand.

So without further ado, let me switch over to our QA/QC demo, so I can show you a little bit about how this works. So in this case, I'm the Head of lending, I'm responsible for loan quality, and I care about reducing the number of bad loans that were originated due to error and staying in regulatory compliance, all while reducing our overhead. And you can see right here, I can view all of the KPIs, metrics that I care about, things like processing time is decreasing. Our defect rate is decreasing, our loan volume is increasing over time with fewer errors. And most tactically up here, I can see all of the top issues that were found during QA review, so that we, as a processing team can improve and catch these issues further upstream.

Now let me switch over to an individual loan where I, as an individual QA analyst would be doing my work. This solution, again, it aggregates all of the data, it stitches together our existing loan origination, core banking and content management systems into one unified view. These systems that were never really designed to talk to each other. And on top, I can see a summary of the loan. So if I come right in, I can see where the loan is at. I can see exactly what my QA agents rather have already done on my behalf. And below are a series of checklists that I have to work in. And before these were all reviewed manually, they were tracked in Excel spreadsheets with dozens of documents and windows open on multiple monitors to triage hundreds of different business rules per loan. And these checklists are now suddenly smart. All of them are processed automatically using UiPath agents, deterministic workflows and intelligent document extraction, allowing me to focus just on the issues that need remediation.

Now let me go into one of these checklists where I see my review is needed. So instead of needing to manually steer and compare between these 2 documents on separate monitors and scroll through them to hunt for what I need. The solution automatically extracts the key data points so I can confirm that they're accurate. So I see right here, the first rule that I need to check is that the name on this document matches what's on this ID. You can see that the agent automatically found that as a match. And I, as a human reviewer can confirm it for auditing purposes. There's a series of rules here that are designed for this specific document.

I see right here that an agent found an issue with one of the rules. And I can see, if I zoom in a little bit closer here that. These dates are expected to be within 30 days of each other, and the agent found that they were in fact not. So clearly, someone entered the wrong date on the credit approval memo. So I, again, as a human can mark this as a no, not matching. But let's just say that I disagree with the agents finding. Maybe you got something wrong, I can easily override that and include a note to indicate why the agent was wrong. And this is both used for auditing purposes, but it also helps the agents learn and improve over time so that it can get more and more accurate.

All right. Let me go back real quick. And I just want to show 2 actions that I can take now as a QA analyst. So one, very often when this is done, I need to escalate back to the processing team so that they can remediate the issues. And before that required me to collate notes on another screen, write an e-mail and send it to them. But I can do all this automatically. I can see right here that the issues have been all summarized for me. These agents know everything about this loan, and I can easily send an e-mail right here. The other action that I typically do is that I generate a report that can be used post closing for auditing purposes. And before this is all done manually typed in a word document, for example. But given, again, the solution has all the necessary context, I can automatically generate this report, which before I again, had to do manually.

So everything you saw here was something that used to take me hours, but can now be done in minutes. And the solution augments and accelerates my entire team of QA analysts. This was built fully as a UiPath process app. On top of Maestro case management, which Raghu talked about a little bit earlier. Everything from application submission to closing, all of the automations, escalation paths, agents are all defined and orchestrated within. That is our QA/QC module for UiPath solution for loan origination. Combined with our loan setup module, they're designed to accelerate time to closing, reduce operating expenses per loan and mitigate bad loan risk, all while working with bank's existing stacks.

Thank you. Raghu, I'll pass it back over to you.

Yes. Mark just showed us what this -- what our overall platform and product look in practice. Now let's bring it home, like why buy UiPath? We opened the session with the reality that CIOs face that about 50% of them struggling with integrations and data quality issues all kinds of governance challenges and shadow AI. These days aren't just AI problems. They're also orchestration problem. And orchestration is exactly where we're building our moat. We are the category leader in task automation that proven at scale and in the most complex and regulated industries.

We're adding coding agent support, lock stock and barrel all throughout our stack to take developers from natural language to production-ready workflows. Obviously, we're building agentic case orchestration and case management with the case manager agent that I talked to you about earlier that coordinates work across people, humans and robots and drive processes, the most complex business processes along. We're delivering out-of-the-box vertical solutions. Just an example of which Mark just showed. And then all of this is built on our enterprise grid governance and trust layers. Now each of these moats reinforces the other. No one else we believe, has this combination, the depth of automation and the breadth of orchestration and the discipline to deliver these as outcomes, not just tools. And that's why we believe we are uniquely positioned to drive a lot of value to our customers upcoming.

Now I want to switch to helping bring this to life with the real customer example. We recently sat down with Jason Paris, the CEO of One New Zealand, one of the country's leading telecommunications providers who is driving a pretty significant agenda to leverage AI as a competitive edge across the business. His organization, we believe, is a good example of what's possible when you combine agents, deterministic automation and orchestration within a single platform. And UiPath is at the core of their transformation strategy.

I think it took 5 weeks to bring their order to cash process into production. It reduced their cycle time, the processing times from multiple days, 4, 5 days to 5, 10 minutes. And they're not scaling their overall B2B operations with expected tens of millions in savings. What stands out is this isn't a one-off use case. It's the platform that they're bedding on for their long-term transformation with orchestration at the center. Let's hear directly from Jason. Jake, do you want to take it away?

Sorry, Jake and Allise can you take it away, please? Appreciate you taking the time to share your story. Can you walk us through your transformation goals and how you see One New Zealand evolves as AI transforms our industry?

Raghu, thanks for having me and also thanks for the partnership that you give us. We've been deploying variations of artificial intelligence for over a decade now, thousands of RPAs in our organization using large language models, generative AI and now Agentic AI. Our goal is to be the most AI-enabled telecommunications company on the planet. And the only way that we can do that is with pace. We're a small market, the bottom of the South Pacific.

And so when we're working with partners like yourselves, the thing that hopefully attracts you to us is the pace with which we will experiment and that we will deploy the technology. We have a secondary kind of mission, which is AI first, but human where it matters. So it's also important to state that AI is going to transform our entire organization but it's not going to stop human-to-human interaction being really, really important. In fact, what we're finding is that it gives us more time to make those human moments even more important. And the way that we can do that is by using a partnership with you to automate its scale. And so there's pretty much -- not a single part of our organization currently, which is not being process mapped, rewired, automated and having agentic tools laid on top of that.

Yes. It's amazing that you've been able to take your employee base along JP, as you've incorporated AI into your technology stack into the way work gets done in One New Zealand. Now I'd be curious to understand what are the key parts of your AI transformation strategy? And then how does a platform like UiPath specifically UiPath Maestro fit into that strategy? And I'd be curious also to learn a little bit about specific impact or ROI that you achieved with the platform?

Yes, that's a great question because I think everyone is deploying artificial intelligence very few being able to bank the cash. That's not the case with our partnership, which is why we are scaling our partnership with UiPath. So as I mentioned before, there's not a part of our organization that we are not trying to process map, automate and rewire using advanced artificial intelligence tools. And that -- an important part of that ecosystem is our partnership with UiPath. Your Maestro tool, we see as an orchestrator over the top of our AI and systems and people. The thing we love about it is that we're a legacy business. We've been around for 20, 30 years. We've grown through acquisitions of different types of businesses. We've got multiple stacks, mobile billing platforms.

And so what we haven't needed to do is a major re-platform or replacement to partner with UiPath. And so the ability for you to map and then automate and orchestrate legacy technology without having to replace it has been awesome. I'll just give you one example. So customer -- a business customer wants to replace the handset either because it's broken or they need to refresh it. That's a path that goes across mobile parts of our businesses using multiple technologies, multiple processes, including external technology and external support. Currently, we are used to be about 4 to 5 days to make that process happen end-to-end. And then it's not acceptable when your mobile phone is your life remote. If you want to refresh it or you need to get it replaced, you need that replaced within a day, not within days. And so what we've been able to do with UiPath is exactly what I've just said before, proceeds may automate, have an orchestration layer over our existing processes, no change to existing processes, and we've changed that 4 to 5 days to 5 to 10 minutes.

How incredible is that where you can use this technology with your existing technology, your existing processes, your existing workflows and move from 5 days to 5 to 10 minutes. So the ROI on that, of course, is extremely strong, and that's why we're scaling this across the organization.

Yes. JP, I mean, your commentary here really resonates. I think the most complex enterprises, such as yourselves, is a combination of modern and legacy technology stacks and our orchestration layer, as you found out and as you know, incorporates the most modern technologies as well as the legacy technologies and brings it all together. So you don't have to rip out what is working for you. You don't have to forcibly modernize what is there for you and so on. So it's great that, that our orchestration platform has worked for you in the way in the way that it has. I also learned, JP, that you went from a proof of concept to production grade deployment in just a handful of weeks, like 4 or 5 weeks. Can you talk a little bit about what enabled that level of speed for you?

Yes. Well, again, we think speed is an advantage for us, not just to attract partners like UiPath to work with us, but also as a differentiator of market. And to be fair, we did have our prior proof of concepts. As I mentioned before, we've deployed thousands of robotic processes across the organization. In fact, we would estimate that we've made about 20% more people in our organization, we've currently got if we didn't have robotic process automation in place. That's a significant competitive advantage and cash advantage just there. The proof of concept that we had with UiPath gave us a huge amount of confidence. It's an integrated platform, AI plus RPA plus orchestration. And again, because that proof of concept worked well, that meant that we wanted to scale quickly.

So I think we built our very agentic agent within about 12 hours, and then it took us a few weeks to deploy it because we've to clean the data up, make sure that it was operating appropriately. And so -- that's why we had so much confidence to scale so quickly. I would say that's not just the mix of the technology, though, Raghu. It's also the subject matter experts, the capability that we've had sitting side by side. We have a kind of 2-in a box model. I think as you'll be aware, we've got our own experts sitting beside your experts working on the issue. So it's something that we've really benefited from and getting UiPath's expertise to upskill and reskill our own people within the organization at the same time.

And then, of course, you want to make sure that you test it end-to-end for scalability. And so again, the proof of concept did that well. 5 weeks, we can see the value and now it's being scaled across the organization.

Yes. The partnership has been incredible across our teams for sure. Now as you -- JP, as you scale this technology across your organization, where do you see the biggest opportunities for you next? And how central is UiPath to that longer-term AI transformation strategy for you?

There is genuinely no part of our organization that is not going to be transformed through this technology. And so your biggest decision is where do you prioritize first. And so we're prioritizing where really the biggest layers of volume and complexity and cost set. So areas like provisioning, finance, risk, fraud and also even really big complex programs in IT like SAP upgrade. So -- but that's just our first bucket of priorities. Genuinely, I can't see any part of our business that's not going to benefit from this -- from Maestro and from the technology you're providing us.

Yes. And we're looking forward to supporting and partnering with you through this transformation that I know lock stock and barrel you're going through at One New Zealand. That's fascinating. Now you've -- I know you've evaluated a number of leading AI and automation platforms at One New Zealand. What ultimately led you to choose UiPath? And then what gives you confidence that UiPath can support that mission-critical execution of your most complex mission-critical automation, it's not just experimentation.

Yes. Well, I think the first part of it is like when you're looking and you're looking at what the technology is available to you, you want to make sure the technology partner is agnostic. And so you need to make sure that it would work across a lease environment with a lot of complex technology. So that was the first tick that the UiPath received. Then also, you want to make sure that it avoids kind of multi-tool complexity. So again, it's an integrated tool that works in combination, not just as an orchestration layer, but across AI and RPA, which makes sense.

And then when you start to test it, you want to make sure that it's got compatibility and you can actually deploy it within your organization, just tick. And then when we did the proof of concept, we could see that not only did it work, but it could scale and you can scale it quickly and you can, as we talked about before, get a cash return on it. So a pretty simple checklist that anyone should be going through as the platform agnostic and can work within your existing environment? Can it be an orchestration layer, which works both with advanced artificial intelligence but also robotic process automation. And then can it scale across the organization and deliver the money step, right, cash that you can either bank or reinvest in other parts of our businesses. So all of those have been ticks for us, and that's why we chose you and we're delighted that we did.

Great. It's always valuable to hear directly from our customers. With that, we'll open it up for Q&A. Unfortunately, we only have time for one question. So I'll combine a couple of themes that we've been seeing come through. Daniel as customers begin to deploy more agents, what are you seeing in practice around the need for orchestration? And more broadly, how do you think about adoption of Agentic solutions and UiPath's right to win in this space as it becomes more crowded?

Yes. Allise, I would like first to give a quick explanation of what's the difference between agent-to-agent, orchestration and process orchestration because I think there is a bit of a confusion in the market. I think when people speak right now about orchestration, I think they implicitly pursue some kind of purely agent-to-agent orchestration, like having a swarm of the agents, we give them a goal. And the agents will communicate to each other, create the planning, will split the task something maybe more akin to like open grow is happening.

When we speak about orchestration, we speak about process orchestration. So it means that in order to achieve an enterprise goal that is being compliant with all the regulation, the regulations in place and understanding the complexity and the many actors involved. You need a bit of a different approach. And typical way to an enterprise solve process orchestration is to have like a process view process description. Many people would use something like this business process modeling notation for showing depicting the process, the workflows involved with the caveat there can be, as I said in the beginning, hundreds of sub workflows there. And each sub workflow can have many steps. Some steps can be purely deterministic and they can be sold by RPA or API, automation, some steps will be agenetic, some steps definitely will require humans in the loop to supervise as you've seen in these demos.

It's kind of clear from all the customers I talked to. In my case, it's almost no exception that the preferred method of bringing AI into the context of an enterprise process is basically injecting AI steps in a deterministic orchestration and workflow engine. In this way, the AI is limited more to understanding a specific task work, understanding the work at the specific stage in the process.

So yes, I would say that the advent of agentic makes even more compelling for enterprises to have a platform that offers a built-in process orchestration. It's much more -- it's much easier and more compelling proposition to have in the context of the same platform, the nondeterministic agent code, the deterministic code and the humans and enterprise workflows that organize that basically manage all the interaction between these actors. You can apply the same on same governance, the same security model, you will have the same audit trails, same observability, more the same analytics across the entire process, end-to-end. This is very valuable for enterprises to be capable of understanding every single interaction that happens in order to deliver go across of an end-to-end process.

Great. Thanks, Daniel. Unfortunately, that brings us to the bottom of the hour. So Daniel, I'll turn it back to you for closing remarks.

Thank you so much, everyone, for staying with us. I hope that this session gives you more clarity of what we are doing. If I have to summarize everything, we are extremely focused on bringing coding agents into the picture. I believe this is going to be a big accelerator into the adoption of our platform.

And I want to finish saying if somehow under the hood, we built this amazing and only platform in the market. We are the only platform that is built right now on the top of a new model workflow engine that is really very good for -- to be used by coding agents. And on the top of this engine, we built business friendly way to describe a process using BPM. And then, we have our proven scalable engine that was capable of delivering for many years, automation escape. And we are talking about hundreds and thousands of automations that run in parallel, concurrent run at a big scale, you need to orchestrate them, to manage them, to have to feed them with data to understand analytics. That's not something that you can build overnight in -- and it requires a lot of deep engineering architecture of thoughts.

And then the third important pillar, we have the task automation capabilities. Basically, we have the capability to integrate with every system out there. The legacy system and model system will continue to coexist for the foreseeable future. And it's so powerful to have all of these components in a platform that offer integrated security and governance. So with this, again, thank you so much for staying with us. We would like to connect in the next couple of weeks with as many as you possible. Thank you.

Greetings, and welcome to UiPath's Fourth Quarter and Full Year 2026 Earnings Conference Call. [Operator Instructions] Please note, this conference is being recorded.

I will now turn the conference over to Allise Furlani Head of Investor Relations. Thank you. You may begin.

Good afternoon, and thank you for joining us today to review UIPath's fourth quarter and full year fiscal 2026 financial results which we announced in our earnings press release issued after the market closed today.

On the call with me are Daniel Dines, Founder and Chief Executive Officer; and Ashim Gupta, Chief Operating and Financial Officer, to deliver our prepared comments and answer questions. Our earnings press release and financial supplemental materials are posted on the UiPath Investor Relations website. These materials include GAAP to non-GAAP reconciliations. We will be discussing non-GAAP metrics on today's call. This afternoon's call includes forward-looking statements regarding our financial guidance for the first quarter and full year fiscal 2027 and our ability to drive and accelerate future growth and operational efficiency and grow our platform, product offerings and market opportunity.

Actual results may differ materially from those expressed in the forward-looking statements due to many factors, and therefore, investors should not place undue reliance on these statements. For a discussion of the material risks and uncertainties that could affect actual results please refer to our annual report on Form 10-K for the year ended January 31, 2025, in our subsequent reports filed with the SEC, including our annual report on Form 10-K for the year ended January 31, 2026, to be filed with the SEC.

Forward-looking statements made on this call reflect our views as of today. We undertake no obligation to update them. I would like to highlight that this webcast is being accompanied by slides. We will post the slides and a copy of our prepared remarks to our Investor Relations website immediately following the conclusion of this call. In addition, please note that all comparisons are year-over-year unless otherwise indicated.

Now I would like to hand the call over to Daniel.

Thank you, Allise. Good afternoon, everyone, and thanks for joining us. I want to start by thanking the people who made this year possible. Our employees, we executed with discipline and purpose. Our customers who trust us with their most critical workflows and our partners who have made a genuine bet on our platform. This is a team effort, and I feel that every day. We delivered another strong quarter, beating the high end of our guidance across all metrics and closing out a year of disciplined execution.

Fourth quarter ARR reached $1.853 billion, up 11% year-over-year, driven by $70 million of net new ARR and the revenue of $481 million, up 14% year-over-year. Alongside that growth, we've achieved full year GAAP profitability for the first time in our company's history. We grew fourth quarter non-GAAP operating income to $150 million, a 31% margin, a reflection of the operational progress we made throughout the year driving meaningful efficiency while continuing to invest in growth. And in Q4, we posted our strongest sequential net additions of customers with $1 million or more in ARR in 2 years with deals over $1 million, up over 50% year-over-year, a reflection of both improved sales execution and deepening enterprise platform adoption.

I have never been more energized. What we are seeing now goes beyond a single quarter, we are at an inflection point in how software is built. Advances in AI are dramatically reducing the time and cost required to create software. And that has led to understandable questions in the market about how value will be created going forward. Historically, moments like this don't eliminate software, they shift where value is captured. Enterprises don't simply pay for quote they pay for trust, for operability and for government, the ability to run complex systems reliably, securely and with full accountability as the cost of building software falls the value of platform that can safely govern, orchestrate and scale that software rises.

And there is a second dynamic that I find even more exciting. When building becomes cheaper, more gets built, more processes get automated, more edge cases get addressed and more systems become autonomous. That expansion does not shrink the need for enterprise orchestration. It increases it. And this is precisely the environment UiPath is designed to operate.

We entered this new genetic era with 4 advantages. First, a unified platform combining deterministic automation, agentic automation and enterprise-grade orchestration with governance, security and scalability built in. This is the full stack, it is what wins new logos and drives expansion across our base; second, a powerful installed-based flywheel, thousands of enterprises run mission-critical workflows on UiPath today. And within those workflows, there are opportunities for agents to be deployed and the overall process to be orchestrated.

Third, 2 decades of enterprise trust and governance, deployment experience that AI plus automation is expected to deliver accountability, ability, observability and reliability at scale; and fourth, the vertical expertise with enterprise-wide reach, regulatory depth in the industries where the stakes are highest paired with the horizontal ability to orchestrate across the entire enterprise.

Let me spend a few minutes on each. [indiscernible] our unified Agentic automation platform. As AI makes intelligence more accessible, what matters is execution. Enterprises are getting answers to complex questions faster than ever before, and yet they still struggle to reliably execute complex cross-system processes with accountability and compliance built in. The goal now is to pay the insight they are getting with the actions and execution that our platform enables financial reporting claims processing, regulatory compliance. This cannot be improvised. They must be institutionalized.

Enterprise automation requires 2 modes, determinist for precision audibility and agentic for reasoning and adaptability. Most vendors offer 1 or the other, UiPath purpose built to integrate both under a single control play, allowing enterprises to move from experimentation to scale production grade deployment. Most people think orchestration means agent to agent coordination. Real enterprise orchestration brings together agentic automation, deterministic automation and humans because that is how work actually gets done. We offer that and the full execution layer underneath it, governing how our transaction moves from start to finish and ensuring that it completes reliably every single time. This is what Maestro is built to do at enterprise scale.

What makes Maestro uniquely powerful is its architecture. It is built on temporal, the most modern workflow technology featuring durable execution and trusted by the most demanding technology companies in the world. Workflows are defined in a way, AI agents can generate and modify it directly while remaining fully transparent to business stakeholders and auditors in a world where agents are increasingly the ones creating and maintaining workflows that distinction matters enormously. The customer results make this concrete, a U.S.-based semiconductor company fail to deploy an agentic workflow with another vendor after more than a year of trying with UiPath, they were successful in under 2 weeks leading to a 7-figure expansion across agent builder, Maestro and Test Cloud.

Today, they run over 3,000 automations and have sales more than 2 million hours and 1 New Zealand who went from proof of concept to production grade pilot in 5 weeks reduce 4- to 5-day order-to-cash process to 10 minutes, and they are now scaling this across their B2B sales operations. With UiPath, they expect at in cost savings this year as they plan to further leverage the platform to support their broader transformation programs.

Driver 2, the fly well inside our installed base. The most important story this quarter is the economic shift underway inside our installed base. Customers are not experimenting with the. They are expanding their operating model on our platform. AI product ARR which includes genetic, IBP and Maestro reached nearly $200 million this quarter with strong growth fueled by Agentic. But the number I keep coming back to is this. The number of customers above 100,000 in ARR, who have bought AI products grew 25% year-over-year, and they spend nearly 3x as much as those who have not.

Additionally, of our top 20 deals this quarter included AI products. All of this is clear evidence that Agentic automation is becoming central to our largest customers' road maps. Importantly, this AI growth is layering on top of a core unattended automation business that continues to grow. We are not seeing agents replacing deterministic unattended automation in production, we are seeing customers extending their processes with AI. A major U.S. airline illustrates this well, building on their deterministic Foundation, they are now deploying agent billers, communications, mining and Maestro to automate, procure-to-pay and supplier workflows a propane for how customers move from task automation to end-to-end process orchestration and how the journey drives platform-wide expansion. This is the fly with, every workflow automated, a new surface area for agents. Every agent deployed drives more automation, deeper integration and broader platform adoption. Testing is another area where we see a significant and underappreciated expansion opportunity as the genetic workflows and applications roll.

Traditional QA simply cannot keep up. Forrester named UiPath, a leader in the Forrester wafer autonomous testing platforms in Q4 2025, with Test Cloud receiving the highest possible scores in 7 criteria, including vision, road map and automation creation, orchestration and execution. A global technology company is a strong example standardizing the entire automation program on UiPath expanding into test Cloud and planning to implement UiPath agents and Maestro to automate supply chain workflows.

Turning to driver 3, governance. Building an agent is becoming easier, making it enterprise grade is not enterprise-grade agents require deterministic execution with traceability in handling and audit trays that satisfy external regulators. We see this play out in how customers choose us. An American Credit Union selected UiPath as we were one of the only solutions to meet their strict banking, security and governance requirements. And a European automobile manufacturer chose UiPath as the foundation of their Agent strategy, selecting Maestro because we could deliver enterprise-grade governance air handling and human in the loop [indiscernible] at the level there compliance standards demand.

In both cases, governance was not a consideration. It was the deciding factor, and that brings us to drive 4 vertical depth. It's not just about governance, it's about knowing the domain deeply enough to manage and operate it at scale for real impact. That is why vertical death matters more in the Agentic [indiscernible] not less. As building becomes easier, differentiation shifts to domain-specific workflow intelligence, especially in industries where the cost of getting it wrong is existential. Advising February, we launched Agent AI solutions purpose-built for health care, targeting revenue cycle management, medical record [indiscernible], claim denial resolution and prior authorization.

In line with that strategy, we acquired WorkFusion in February, bringing purpose-built agents for financial crime compliance with deep anti-money laundering and know your customer exported directly into our platform, expanding our reach into the highest takes compliance workflows inside global banks. Health care and financial services are 2 examples of a broader strategy. We pair vertical depth with the horizontal reach to orchestrate across every function of a global enterprise, a combination that either horizontal or vertical platform alone can match. And great platforms don't scale along.

Our partners are building practices, joint solutions and go-to-market motions around our platform. Our expanded partnership with Deloitte is a strong example. Together, we launched Agentic-ERP embedding AI agents into mission-critical finance and operations workflows a Fortune 20 oil and gas company that is migrating to SAP S/4HANA is already scaling through the partnership, expanding test cloud coverage from 10% to roughly 50% on their SAP environment while building new Agentic use cases across the migration.

Accenture tells a similar story. Together, we deployed the global agentic sales order entry solution for a strategic life sciences customer, reducing processing time by 1/3 unlocking automation for orders previously too complex to handle and orchestrating autonomous agents transforming the orders while navigating 150,000 exceptions. [indiscernible], I want to give you a preview of what's coming next on our product road map. Over the last few months, the world has changed. The boundaries of what is possible have shifted faster than most people expected. We have spent years building a unified platform for exactly this moment. And what it can now unlock with the next generation of coding agents, it's something I'm generally excited about.

Our platform is evolving into 1 where coding agents can participate across the entire automation life cycle. Agents will work with subject matter experts to discover processes and identify exceptions. They will work with business analysts to generate process definitions. Since developers in building automation, deploy those automations into production and help manage them at [indiscernible] The first capability of that vision ships in the next couple of months and it targets a problem I hear in nearly every customer conversation. Their automation backlog is growing faster than their ability to build. The ROI exists the executive...

[indiscernible]

Hello, and thank you for your patience. I will now hand the call over to Daniel Dines.

Hello, everyone. Thank you for coming back after our outage with the service provider for our Investor Relations conference calls. We are ready to take questions. I hope that you guys get the chance to listen to the end of our reading. And also, we have published online the entire transcript of our earnings calls.

So thank you again and apologize for the delay. We are ready to take questions.

[Operator Instructions] And our first question comes from the line of Bryan Bergin with TD Cowen.

First one I have is just as it relates to net new ARR. And as you build the 2027 outlook -- just how are you thinking about net new ARR expansion potential here on an FX-neutral basis? Sorry if I missed what you said on FX contribution assumptions as it relates to 1Q and the full year. But just trying to unpack that looking ahead, and then my follow-up is going to be on margins.

So an op income margin, I appreciate the update on the 30% target. Just want to dig in on how you're thinking about the potential kind of the moving parts of that as it relates to gross margin and OpEx components moving forward?

Yes. So Brian, great to hear from you. When you think about the IRR contribution, I think our guidance kind of says that there's really no significant or material FX contribution from that versus our prior guidance. So as you look at it, really, FX is a minimal impact from where our previous estimates were.

The second piece of it is, from a margin standpoint, you look at the moving pieces and definitely across the board, there is opportunity to identify and to use the technology advances across every function. That includes engineering, G&A as well as sales and marketing, which gives us really the ability to continue to reinvest in growth as needed. But we're going to look at a balanced way in terms of what makes sense for the company. And you can see our commitment to operating margin expansion over the last two years.

And then just back to the IRR, I want to just give a little bit of color. When you look at our base, we have a sizable Japan business. So we have headwind from the yen, and tailwind to the euro, and they basically net out to be an immaterial impact for the full year. So we're really pleased with the progress. As Daniel commented and I did in the script, we really feel positive about the expansion that we're seeing within our customers and our ability to stabilize our net new ARR, and that's kind of reflected in both our performance as well as our guidance.

And our next question comes from the line of Sanjit Singh with Morgan Stanley.

Daniel, thank you for the disclosure on the ARR traction. I'm sorry, the AI traction with respect to ARR, the $200 million that was great to see. In terms of the composition of that, could you give us any details on sort of the split between IDP and what you're seeing on the Asian side. And to the extent you can't sort of disclose that, I'd just love to hear about the underlying momentum with the agentic side of the house, including Maestro as you go into next year?

Sandeep, we have really a great momentum on diffusion of the AI within our platform. We have not provided clear ratios between different components of what we put into the AI. And I will let Ashim to comment further.

So like when you look at the way we price, we actually allow pretty good fungibility between our AI and Agentic products actually, both in some of our old pricing as well as our new pricing. So we don't really materially split it out.

Of course, IDP hasn't been in the market for a longer period of time for like the last 2.5 to 3 years. So IDP definitely has a good portion of the IRR. But genetic is a significant portion, and we see that in the platform. You can see that in the deals and the commentary that we're giving and selling as a part that we talk about in our script. So from that standpoint, we can't really split it apart -- but we also see them as complementary because remember, IDP also includes XP, which is not like simple document processing. It really uses advanced technology to be able to parse different documents using different models and that is part and parcel of the way we price.

Yes, that's great. That's great context. And then just a follow-up on the guide, time on 2 aspects. One, in terms of work Fusion, how should I think about that contribution I sort of calculate what the guidance implies from a net ARR basis. I think there are some reports out there that they are around $25 million ARR toward the end of last year. So I just want to sort of stand to check that. And then from this time last year, there was some concerns around do as you guys are pretty cautious on the federal business, just for sort of underlying assumptions about Fed going into next year, maybe the first half of this year given some of the headwinds you saw this time last year?

Yes. So the first thing is the $25 million is not accurate. That's the first thing I can say categorically. The second piece is, they also had a different method of accounting. So when we brought it back, even the numbers that have been out there also do not account for it. It is actually below our materiality threshold, Sanjiv. So that gives you an indication. We really look at this like a tuck-in acquisition in terms of where it is. And from that standpoint, you can also just see kind of the strength overall within our guidance, and we've been transparent that, that includes the WorkFusion contribution, but it is immaterial, and we don't break it out.

And then just on the Fed piece?

Yes, sorry. On the federal government, we're actually seeing a really good traction there. I would say just like the environment, I would say the federal government is a dynamic economy. But I would say our team has done an incredible job connected at really high levels within the organization. And I'll let Daniel comment on some of his discussions and his views of it. But within certain agencies, we feel very well strong position. And then there are some agencies, of course, that are going through their changes. But overall, we're actually very bullish about the way our teams are executing and the opportunity that exists there.

Yes. And we are seeing an increased appetite for more long-term projects, strategic projects, especially in the department of world.

And our next question comes from the line of Michael Turrin with Wells Fargo Securities.

Just to start, maybe a higher level one. You had some commentary, but just in terms of budgets and what you're seeing around categories like automation, in AI, it would be great to get just a top-down view there and also how you're positioned to capture that in the market where there's just an increasing number of vendors also positioning agentic solutions, which may be newer to market, but might also insert some noise into those conversations.

Okay. I think we are really well positioned to help customers with the diffusions of AI within their enterprise workflows. We are -- we have -- we built Maestro, which is essentially a process orchestration technologies that -- and at its core, is a new powerful workflow engines. That's -- that gives us a very interesting advantage in the market right now.

So we all know about the impact of the coding agents. I would say that this will translate for us. And I'm extremely bullish about it. into a much faster adoption curve for our customers. We aim to use coding agents to enable our platform for coding agents that will accelerate dramatically the time to value for our customers. And that, of course, includes creation of AI agents, deployment of agents in the context of enterprise workflows. I would like also to stress how important is the combination between deterministic automation and Agent automation into the context of the same platform that can orchestrate both what I would say, humans, agentic and deterministic automation.

Ashim, just you gave some texture. I know the commentary and the guidance on the call was pretty similar to entering fiscal '27, '26, but it sounded like in some of the prior answer that maybe public sector is trending a bit better. So just any more context you'd give us around how you're characterizing the current environment the visibility you have into the model for the forward year at this point and just how you're thinking about the contribution from the AI product portfolio as that scales in fiscal '27?

Yes. I mean we really continue to characterize it as variable. And I'll double-click just again for anybody who's new in terms of what -- but I think we do see pockets of strength and we see pockets of pressure or fluctuations that happen from a macroeconomic standpoint. And at the same time, those tend to move around quite a bit. Like right now, our bullishness in terms of public sector feels really good.

Last time on this year, if you remember, we kind of felt a lot of uncertainty in that area. We're seeing strength in areas like financial services and health care, international markets like Australia. And then there's -- obviously, the Middle East conflict is there, so there's uncertainty there. So we really characterize it as variable. As I commented in the script, we continue to kind of maintain a very consistent guidance philosophy. We look at our pipeline. We have really deep inspection. We get a lot of signal from the field. Daniel has spent a lot of time with customers over the last 3 months, 4 months. We have -- we've been very in touch with kind of the field in terms of hearing. And then the other piece is, we obviously have a very strong now statistical and forecasting models between our finance and our ops team, and we triangulate the 3 of them. So we talked about kind of putting the appropriate prudence in the -- for guidance, accounting for the variability in macroeconomic environment, and we've done so. And at the same time, when you look at our guidance, I do think it also reflects kind of stabilization of net new ARR and what the potential is yielding in terms of the traction our teams are making in the Agentic market and how we're positioned. So that's how I would characterize it.

And our next question comes from the line of Kirk Materne with Evercore ISI.

This is Chirag on for Kirk. You highlighted multiple industry partnerships, right, Veeva like with Veeva and certain vertical solutions like health care and financial crime, would you highlight health care and finance as the 2 verticals that are showing the strongest willingness to spend right now on a genetic AI initiatives? Or -- are there others that you would flag? And when you think about genetic automation at scale, what does success look like in terms of repeatable playbook and sales cycle impact here?

I think you got it very right. It's the health care. And I think we nominate it within the health care in particularly, I would say, parts of revenue cycle management, denials, prior authorization. It's a very important type of processes for us. Financial industry has been since the beginning of the company, our stronghold, and we strengthened it with the acquisition of World Fusion with our big foray into financial crimes. And I would add also the public sector is an important vertical for us that we are eyeing.

And our next question comes from the line of Terry Tillman with Truth Securities.

I have two. So first on Maestro is my impression, it's vendor agnostic from an gentex standpoint. Are you all seeing situations where it's involved in managing agents from system or record companies or AI-native businesses? Or is it mostly like a control planning for your own agents? And then I have a follow-up.

Yes. I think Maestro, it's kind of agnostic in terms of what kind of agents it can manage. Of course, for our own agents, that are built with agent builder, we have very tight integrations. But we have also brought agents built with open source frameworks like the land graph type of agencies, first-class citizens in our platforms. And in terms of using -- utilizing agents built on system of record applications.

Of course, we facilitate using them in our platform. I would not say we manage them. It's more or less like you can call an API that is provided by that platform. But I want to be specific, the all agents that are built with open source framework can be deployed and executed in the contract security and governance that our platform provides.

Yes. That's a good clarification on the API side. Thank you, Daniel. And I guess, Ashim, the SaaS shift, that was an important call-out, 1% impact to growth as we look into FY '27. I'm also curious though, is there also starting to be this impact of timing dynamic or around consumption or scaling volumes related to the actual agenetic solutions that we need to kind of appreciate that's not going to show up in revenue yet.

No. I mean, remember, we do -- we still price on kind of a bundle, meaning on a subscription, consumable-type hybrid model, meaning we sell kind of use it or lose it units that are there. So we're not on a consumption basis of accounting, so to speak. We're still on an ARR basis of our accounting.

So I would say there's -- it's not about any trailing or any delayed impact that you would see there. At the same time, I think our [ Gentex ] solutions are scaling and our customers are adopting more and more as we talked about in the script and sales are moving very well for us. And that obviously is what's contributing a little bit to ourselves.

And our next question comes from the line of Radi Sultan with UBS.

Daniel, in your prepared remarks, you mentioned this growing backlog of automations you're seeing at customers. I just wanted to double click on that, like how big is that tailwind of AI unlocking more automatable workflows. And you mentioned the AI product there, but just how material is that sort of pull through to the core automation business as well? I just love to get your thoughts there.

Yes, that's an acute observation. Because of the huge interest in AI, it's actually driving renewed interest in automation. I think in most cases that we are seeing, people expect that the use of AI will result in some sort of automation. And it's becoming more clear that AI and Agnetic AI and the terminating automation are very complementary.

So basically, any AI initiatives surfaces more opportunities for deterministic automation, especially in our case for unattended deterministic automations.

Got it. And then just a follow-up for Ashim. Just as you think about the ARR and revenue guide for the year and we think about sort of what the biggest drivers are you guys really extended the product portfolio over the past 12 to 18 months. And just as we think about AI product, test cloud vertical solutions, sort of core RPA. Like how should we think about sort of what the biggest drivers are of that sort of growth next year as you kind of think about the guide?

Yes. I think if you just look at some of the metrics that we disclosed, right, 90% of our $1 million-plus customers haven't incorporated AI products, right? I think that is a great -- to me, kind of a great tell of the success of the AI products and the ability for us to expand. And we've also talked about the number of customers that still have room to adopt those AI products that are there.

So from our standpoint, AI and agentic is going to lead the way. But at the same time, as Daniel talks about, they're not a separate stream. They actually are very synergistic. As people pull forward AI and agentic products from us, it actually also pulls through the rest of the platform, whether that is IDP IXP unattended robots, et cetera. And we see that. We are very purposeful in discussing that we are seeing growth rate within kind of the core RPA business and we look at that as very synergistic as we go forward. The other thing to highlight is we're super excited about our test automation business. And that is still in its infancy, but we really see that having good traction in the market, and that can also be -- that is also a growth driver for us as we enter this year.

Our next question comes from the line of Scott Berg with Needham & Company.

I've got 2. Daniel, we've been doing some more question partners here. It's become very evident and clear that your partner strategy seems to be resonating really well right now across several different or your vertical strategy, excuse me, is working well across several verticals. But my question is, as you look -- are you able to lean into that strategy even more so given the success you're having there lately? Or do you feel like you're already at kind of a maximum effort.

On the contrary, I think we are at the beginning of our vertical strategy. We are doubling down our focus on investments into this year. So if I can summarize our product strategy, I think there are 3 major pillars that we are seeing right now. So we focus on adopting coding agents all across our platform.

So every single artifact is building on our platform will be built primarily by coding agents. Second its process orchestration that really drives everything Agentic AI and deterministic workflows. And third, it's vertical solutions. And we have seen clearly more of a move into customers that have a higher demand of kind of an outcome-based vision by use case-based type of solutions that they want to adopt.

Got it. Very helpful there. And then Ashim, I was hoping you can drill down in the quarter a little bit I know there's a $14 million tailwind around FX for ARR. But what was your assumption of that number going in the quarter? Get a lot of questions to try to kind of back into the math in terms of how much incremental impact they might have been versus your expectations 90 days ago?

Yes, it was honestly right. It was just right in line with that. As I talked about, like I think the yen you could see has an inverse correlation to the euro and the net for both of those tended to be zero. We see that both as we look into the current year as we've seen FX rates move as well as the current assumption that we see there. So from both our guidance standpoint and our results, we really see an immaterial impact to that. The driver for our beat in the quarter was really just sales execution. And we're -- we feel very strong about the customer response as we've seen about the traction that we're getting within our AI products. FX did not have a material impact versus our guidance.

Our next question comes from Kingsley Crane with Canaccord Genuity.

And I think the idea of AI on top of deterministic automations, is really resonating. Just on this idea of Agentic really being about pulling through to the whole platform. Just trying to get a sense of how that ends up playing out from a deal timing perspective? Like -- is the customer typically renewing at a much higher rate? Is it happening where they'll adopt AI and then through the life cycle of their contract they'll realize that they need more automation? Just trying to get more color on that.

I think it's all of the above. Honestly, like we've seen the customers renew just at renewal, expand into AI products. We have very good examples of that, both within -- across every vertical and every geography. There's also areas that they're still working through their POCs, but it's bolstered their renewal and their confidence given our road map. And the POCs are moving well, so they would expand just a little bit as they continue to kind of dip their toe in the water.

So from our standpoint, it's not one single motion. It really depends on the customer or the circumstance. But what is encouraging to us is the success that our proof of concepts the feedback that we're getting from customers that as Daniel talked about governance matters and the full extent of our platform is a difference maker for us.

Great. And then just a quick follow-up. That #1 OS fold ranking for screen agent definitely impressive, and that's still holding up -- just curious like how specifically Screen Agent is driving more automation growth within customers. And just a reminder on the unit economics that's affected by running OPIS versus running HiQ, things like that.

Yes. I think we are still in the early innings of deployment of the screenplay agent. We are seeing really good use cases from our customers. They -- the powerful use of this screen play agent is that it is used in the context of autonomous workflows.

So basically, the best we combine like using the terminating UI automation technologies. And in the places where it extremely difficult to define in rules how to use the screen when the screens are -- have a high degree of variability.

Our customers are using the screenplay agent. So that basically extended our plus 4 in a few use cases that we couldn't basically touch before. But again, I think it's still early to comment on how does it help with the platform adoption.

And our next question comes from the line of Arsenije Matovic with Wolf Research.

I just kind of wanted to go back and expand kind of on the ARR guidance methodology in terms of that conservatism. Like what does that mean? And I understand we're not going to be talking about inorganic from WorkFusion, $20 million, whatever it is. Even if you strip out that number growing at the 65% rate the CEO talked about. Is there a way that it still looks a little bit less conservative in that guide? And if there is a little bit less conservative a dynamic where it's just, hey, larger renewal cohorts and also more confidence in that execution tailwind that you started to see exiting the year?

Yes. So one is I just want to correct, Like, I don't think we should -- the metrics that we talked about, as I said, we bring it on at a different ARR methodology. So I really want to caution everybody to use kind of those -- those assumptions. It's immaterial for a reason as we've done that test. The second piece is, while the business was growing at 65%, remember, we also have overlapping customers, et cetera. We really view this as a technology tuck-in that can drive utilization and stickiness across our Agentic and AI platform. And of course, we do see potential there for the upsell, but we also have to go through an integration period with the company. And that is all baked into our guidance from that standpoint. We look at it as our core business continues to be very strong, and we are stabilizing net new ARR. And with AI and Agentic, we do feel bullishness about the overall business. But given the macroeconomic environment continuing to be variable, we do layer the appropriate prudence that is there.

Got it. And then just in response to an earlier question, I didn't really kind of get the in line with the constant currency. Can we just clarify what was the constant currency ARR growth rate implied in the guide for revenue and for ARR growth because the communications are out the year on tailwinds and incremental headwinds has kind of laid up a weird kind of analysis to figure out what the actual core constant currency growth was?

Yes. From our standpoint, we gave the $14 million, which we assumed -- which we -- for the guidance that was there, but the growth rate remains 11% for us. It is largely a material year-over-year.

And our next question comes from the line of Siti Panigrahi with Mizuho Securities.

This is Phil on for Siti. So you guys raised the long-term non-GAAP operating margin target to 30%, which is a meaningful step up. Can you walk us through what gives you confidence in that number? And what is the time frame of achieving that target?

Yes. So right now, we're in and around 23% north of that. We've shown really good progress and scalability over the last couple of years, in particular. The first thing is we just continue to operate with really good discipline. And so we constantly are moving investments to higher return areas.

And so when you're able to do that, it obviously creates a scalability of expansion. The second is we believe in the productivity that is being unlocked right now with Agentic and that identification within our own business is something that is very exciting for us and our teams to unlock further steps of productivity. And that includes all areas within the company. We can be more productive, expand and support our broader road map, really with similar technology spend just because of the advances that are there or R&D spend.

The same goes with our G&A function as well as our sales and marketing function. So we're really seeing that scalability just even with the technology advances as well. In terms of time frame, it's a long-term margin target. We -- as it implies, that's kind of within a 3-year time frame from our standpoint in and around it. And at the same time, like we don't take -- we're not waiting for 3 years. We're going to continue to execute and drive productivity as we see fit.

And our next question comes from the line of Koji Ikeda with Bank of America.

I'm going to ask one on dollar-based net revenue retention. So it's down 1 point to 106% when adjusting for FX. And so looking into fiscal '27, what are the main drivers we should be thinking about, whether that's product, geography, vertical or maybe something else in there that can drive expansion in that metric? And how should we be thinking about the dollar-based net revenue retention assumptions that are embedded in the guide? Is that flat, up or down from the 106%?

Yes. I think when you look at overall net new ARR stabilizing, like we don't really see a difference in the mix shift between net new logos as well as expansion. We see them both as areas that will continue. We've kind of operated in this 80/20 70-30 split. So that gives you, I think, enough data to be able to see that net new ARR stabilizes over this period of time from where we are. In terms of what gives us confidence or kind of how we see that expansion, again, as we spoke about earlier, it is really around our AI and Agentic products. And then with that, really pulling through the overall platform, including deterministic automation, continuing to expand across our customer base.

Our next question comes from the line of James Kisner with Water Tower Research.

I guess first, just -- from the foundational model perspective, I mean has the entropic supply chain the designation -- have you seen any kind of ripple from that at all? Is there any kind of exposure at all any change in behaviors out there -- and then just on the Work Fusion acquisition, does that portend potentially future acquisitions and other verticals for identicabilities?

Yes. In relation to Antelope -- our strategy was from the beginning to be model agnostic. And we -- 1 of the features that many of our customers have requested this to give them the capabilities of choosing what model and even bring their own model to be used by our platform. So we do offer entropic models but they are optional and not mandatory. And from this perspective, there is zero impact on our working relationship with public agencies in the U.S. of our Work Fusion.

Yes, it's -- we are always looking into the market, especially for tuck-in acquisition that gives us the talent technology and expertise in a particular vertical.

And with that, ladies and gentlemen, that does conclude the question-and-answer session. I would now like to turn the floor back to management for any closing remarks.

Well, thank you so much for listening to this call. And once again, I would like to apologize for the outage that we experienced, and I'm looking forward to meeting many of you in the coming days. Thank you.

Thank you. And with that, ladies and gentlemen, this does conclude today's teleconference. We thank you for your participation, and you may now disconnect at this time, and have a wonderful rest of your day.

Thanks, everyone, for joining us today. My name is Scott Berg. I lead the enterprise software and SaaS research efforts here at Needham. Thanks for joining us for our 20th Annual Growth Conference here. Today, with us right now, we have UiPath. We have the company's CFO and COO, Ashim Gupta. Thanks for joining us so much, Ashim. Appreciate the time. Yes, got lots of stuff to talk about here. But I guess, for those that are less familiar, how about an overview of UiPath?

That's awesome. So UiPath, founded by Daniel Dines, founded in Romania. As he would tell you, 8 people sitting in an apartment, started kind of with really humble beginnings, never really thought it would grow to where it has been. Around 2015, product market fit. And really, it found its roots in core RPA, so robotic process automation and really through -- really years of research of improving computer models, which allowed it to scale. So it was third -- it kind of entered the market third or fourth back then, and it scaled pretty dramatically. And really, when you look at every quadrant in that category, UiPath was #1 and really had robust global growth from 2015 to where we are today.

We're $1.5 billion -- $1.8 billion plus of ARR. And when you look at it, really global, 50% of our revenue is international, 50% is domestic across every single industry. And really at the core of what UiPath does is we emulate what people do, and to automate and transform processes. The core, as I mentioned, was RPA, but recently, the last 3 to 5 years, we really have scaled in AI and multiple vectors of AI. Advanced intelligent document processing, process intelligence. Those are just 2 examples. And then, of course, 18 months ago, at our Ford, we launched the Agentic vision. It wasn't something that we did in reaction. We didn't repackage. This is something that Daniel really had a vision of starting in 2022 when he coined it semantic automation, really bringing natural language into the forefront about how you develop and how you approach automation and transformation.

And today, we look at agentic automation as really -- and process orchestration as really new tips of the spear that give us really a great view in terms of further driving growth across the world. And then just financially, $1.5 billion of cash sitting in the bank, no debt, really strong performance on our buyback. We bought $800 million plus of stock over the last 2 years back from the Street, returning it to shareholders. And frankly, we're really happy with the progress on operating margins as well, getting to 20% plus, both on free cash flow and operating margins.

All right. So let's talk about products. This has been a fascinating space from my seat to actually kind of watch this evolve over the last couple of years is, there's been a significant amount of investor chatter about how generative AI or automation in general will overshadow or replace what some investors might think is an existing RPA and RPA automation, right? But you had me join you at your partner conference in this fall on stage. It was interesting. Ashim got to interview me instead of the other way around on a couple of things. But I got the chance to speak to several hundred partners. I appreciate all the partners that LinkedIn me for connections. That was great. But you had me discuss what we, from our viewpoint, see in the agentic automation space.

And what was really reassuring is all the partners I spoke with that had a similar view. I've been on the side of I think you all are going to win in this space. The market obviously has not seen that the last couple of years in general, right? But I always thought this was going to be something that will evolve over time. And whether it's orchestration across processes or departments, et cetera, there's lots of opportunity here. But the partners were really reverberating that. I guess product and maybe understanding in budgetary timing is maybe what was missing in this investor viewpoint. But the market's understanding is, I think, starting to improve, stock is working a little bit better, maybe budgets are starting to unlock a little bit. But what's the benefit really of offering this entire expanded solution and platform together? Because I don't think really investors understand how you're marrying all this into a single solution.

Yes, it's a great question. And we were super fortunate to have you, Scott, at our conference. I just want everybody to understand like -- start with the outcome. What is every company trying to do? They're trying to become more efficient. And to become more efficient, you got to transform the entire process, right? The more parts of a process you can automate, the more efficient you get. It's very simple. Break down any process that you know, right? Invoice to cash, procure to pay within your own sphere, right, even processing and analyzing quarterly reports of -- for an enterprise. There is a set of tasks that are -- there are a set of steps that are hugely repetitive, deterministic in nature, rules-based, right? There are tasks that require a reasoning power, right? Like, hey, which are the first 3 companies I should work through? Which companies have the largest variances that have the most potential and you can have a reasoning model kind of work through that.

So when you combine deterministic and agentic automation capabilities, that allows for the most breadth of a process to be transformed. And so when you approach a CIO of a health care company and talk about claims processing, whether you approach the government of the United States and talk about getting more efficient in managing operational data of the Air Force or whether you go to an oil and gas customer talking about procure to pay and all of the kind of the operations that they're doing, we now have a broad platform that processes -- that helps automate the processing of documents, deterministic or rules-based steps and probabilistic steps.

And then you elevate one level higher for process orchestration, which is part of our solution now or part of our platform. Now you have a process and you want to say, how do I observe that process? How do I manage that process? And process orchestration is not about giving a few governance details. It's about creating an entire framework of governance and observability around your processes that gives insights and data, which is hugely valuable to an enterprise, especially in the advent of agentic.

Yes. The orchestration parts like that whole framework is, I think, what investors seem to get a lot -- or forget a lot kind of through that process. Within this framework, you all have announced a string of partnerships with NVIDIA, Google, Microsoft, Snowflake, OpenAI. I think they're all kind of known in this agentic space right now, last time I checked. But what is the advantage of an open ecosystem that you're developing with these partners? And how do these partnerships really enhance what your product is doing today?

Yes. So the second one is going to be bespoke depending, but let me just talk about the first one. The first one is we believe -- I think Daniel is very philosophical. He would tell you he's always believed in freedom as a pillar -- cornerstone, whether it's organization or product. So in our minds, no customer wants to be locked into a finite set of vendors. And if you think about it, 5 -- 10 years ago, there is a little bit of risk to doing that. Now if you're a customer, there's a huge amount of risk. Who knows what's going to happen in the Gemini, GPT, Claude kind of race that is out there. So I think the first piece is having openness of architecture really is about putting the customer ahead of your own philosophies, letting customers have the most amount of choice that's there.

I also think in the world of security, that's super important. We met with one of the top 5 banks, their CTO. They're only allowing certain LLMs to come in and be used across their environment. What if you only partner with one and the other one is left out, right? So I think having choice really matters and that horizontal openness of architecture is hugely important. In terms of the partnerships that you mentioned, they're all very different. If you look at NVIDIA, if you look at Anthropic, if you look at OpenAI, really having -- giving access to the latest models and being able to integrate their models and give our customers fast access to those things as they're developing agents, as they're building automations is hugely valuable.

And some of that is from the customers' own requests, right, because they're forming their own partnerships, their own convictions. So it again brings those choice. When you look at Snowflake, Snowflake and integrating kind of being able to integrate their data know-how allows us to create a data fabric. So if you think about it, we're a zero copy company. So we don't go in and just the invoice data of a customer, and they want control over that. So for us to be able to integrate with Snowflake, where they can spin up for particular processes, specific data sets and store them, but under their own control, that is a great partnership with Snowflake that is there. So each of them has a purpose, but going across the platform is super important and across the ecosystem to give customers choice.

I'm glad you think Needham is a top 5 investment bank. We weren't the customer. I'm just kidding about that. Excellent. So I guess as generative AI drives more complicated use cases, how do you think about the verticalization of the product or verticalization in general, both on the product side and on the partner side because that's an area of emphasis recently.

Yes, it's a great question. The first thing is I think vertical -- I think time to value is maybe the most underestimated part of the ROI equation for customers, right? So we talk about ROI, but how fast they get that ROI is super important. Super important in the selling process from us to a customer, super important for their internal selling to their stakeholders. And frankly, in between purchase of software and realization of ROI is the most -- is the largest period of risk for any software company that exists. So verticalization accelerates time to value. Let me give an example for verticalization, being able to use AI and agentic capabilities to optimize pricing and inventory and being able to automate the scheduling of production planning that happens.

UiPath has acquired that capability from PEAK. Another example is revenue cycle management. Many people have used our horizontal platform to automate revenue cycle management or claims denials within health care. Most every health care company does that in some shape or form with UiPath, like in our customer base. Imagine instead of delivering them the LEGO blocks, we can deliver them a solution that they can plug and play faster that accelerates that time to value. And so when you add that up, I think the time to value is there, but it also expands TAM. Because when you're looking at it, horizontal technology is good, but a lot of times when people are looking for those solutions and say, "Hey, I really want an Ariba, right, to give you an example or I want something that processes procure-to-pay. They think Coupa and procure-to-pay and Ariba, right, just as 2 examples of it.

Tomorrow, if you're talking about how to automate procure-to-pay, we can do a procure-to-pay automation that integrates with the hundreds of platforms that are out there. The last point I would just make about this, Scott, is verticalization is very synergistic to horizontal. So if I build a really strong horizontal base, I can verticalize and build large towers across that foundation, compared to if I put a bunch of towers together and try to become a horizontal leader. So what we really feel is our horizontal strength allows us to productize the large scalable vertical use cases that our customers are asking for.

Okay. How much -- as your customers are going down this journey right now, how much of their use case of the agentic theme has a human in the loop in it, do you think? Is that still a significant component of how they're deploying today? And as they get more comfortable maybe removing that human, does that, I guess, maybe improve or add to your opportunity with these customers?

It's an awesome question. Within UiPath, and Hitesh Ramani, our Chief Accounting Officer and Deputy CFO, is sitting in the audience, we -- there needs to be a proof period, especially for high regulated processes, and you want that human-in-the-loop assurance or that throughput. So it also depends on like where the agent is proving. So if I look at the quality of an agent, if I have an agent that has a 70% -- we score all of our agents, which is a good differentiator in my mind for UiPath is the way we provide scoring of an agent, how accurate they are, kind of the degree of variation and the outcomes, et cetera.

If you have an agent that's scoring 70%, that's still productivity, but it's not enough to be unattended productivity. So we do look at human in the loop to be there. In cases where -- if I'm doing an SEC filing, I don't think there's ever a day where I'm going to let an agent go and put an SEC filing in if there's a 1% chance for an inaccuracy. So I probably will have unattended or human in the loop or some level of control there all the time. At the same time, if I am approving purchase orders against budgets for my engineering team to purchase monitors, I can picture that if I'm 99% accurate, that's good enough, right, depending on the level of risk, both for the enterprise and as well as for the company.

So in our minds, I think, to answer your question directly, there are some processes that I don't think will ever be fully unattended. And I think there's going to be a lot of processes that evolve to be unattended. The more unattended a process, theoretically, the higher the ROI, right? But to do that, I also think we have to continue to be able to improve and integrate and become deeper into those processes. And I think verticalization helps because you start specializing into the systems, the data, the flows, the capabilities of a company. And the more expertise we can bring in our platform, the higher that agent score becomes. So for really complicated processes, we can get to that score where there's more reliability. And I think that is part of our R&D road map.

I think the human in the loop parts, like it's kind of a fascinating item to all this because I think customers want fully agentic processes. But to your point, it's just not realistic because you have to guard against that 0.5% or the 1% issue sometimes at times. But the more you can get there, obviously, the better for what they do. So let's talk a little bit about changes in your go-to-market in the last quarter.

The company just reported its first quarter of net new ARR growth on a year-over-year basis in 2 years. I guess you all have spoken about a lot of go-to-market changes last year with the annual kind of coming back into the role. How much of the improvement there and then your guidance for the fourth quarter, which I think caught a lot of us by surprise on the positive side, how much of that is driven by some of the go-to-market changes or maybe just a change in your customer demand patterns?

I would say it's 80%, 90% execution. I think just the way that we're executing, the stability of the company, the clarity by which we're operating, the principle of kind of no need for large middle management layers, I feel like that's really been a lot around the performance of and -- I would say, predictability and performance of this past year. So entering the year, there was a ton of volatility. I was just commenting to you like last year sitting in this room, we had political administration changes. And if we left that room, I remember there was such a bullishness and then we go in and there's tons of volatility that happened throughout the year, right?

And I think if we look at kind of how UiPath navigated and how we communicated there, I think that's because we are close to our customer, close to our field. And that makes a really big difference from where we are. And I would say the execution is around there. Like being -- we're talking right now about my last 2 weeks, I'd say 40% of my meetings are about first and second quarter, right, meaning getting ahead of that, starting to understand it, starting to look at your top renewals, look at your top accounts. And when you do that, looking at your day 2 implementations. When you're there, it's not all pretty. It's not all pretty, but the longer time that you have, the more that you can affect change. So I think we're just a more disciplined, more stable company, and that's really driving to me 90% of what you're seeing right now.

One of the things that's been interesting for me to track over the last year is the consistency in your messaging around that. I host a bus tour, software bus tour in New York every June, and you all, of course, were on it, both you and Daniel were on it. And that consistency around the changes and improvements and what you're seeing really in the last 6 to 9 months, the messaging has been consistent, which is great. I think the natural follow-up question though on where the benefit is come from is now that you've had the improvements, is it repeatable going forward? Because 2 years ago, it looked like some changes were coming that people -- or that the company seemed to enjoy and it wasn't repeatable. Is this repeatable that it's not just a 1- or 2-quarter benefit.

So I'll say the popular answer -- or the obvious answer, and then I'll try to double-click so it's more interesting for everybody. The answer is, of course, it's repeatable from where our conviction is. It's substantive. It is substantive in the rhythms, the organization, the metrics that we are seeing, right? The more interesting question is it's only repeatable if we continue to not be satisfied with where we are. So if I'm sitting down here in 6 months or 9 months or Daniel is sitting here, I hope it is also not articulating that we're just looking at renewals 6 months ahead, but our account planning is moving to be deeper and longer. Just to give you an example, I think there's opportunity in every company, right? The problem you get into is where you get satisfied with your first step, and the environment changes.

And what was strong in one environment is not strong in another environment. And so for us, I can tell you me personally, I live in a constant state of demand more. This isn't good enough because tomorrow, we just don't know what's good enough. And I think that is what we look at every single day, right? So I love that we're going ahead on our renewal planning that we're there. I'm super happy right now about the level of project management that we're putting into some of our installations and our services installations that are there. I'm super happy about Daniel being on tour with certain customers that have had bad experiences in the last 2 years, not just the ones that are good that we have deals in front of us right now. And I think those are the things that give confidence for me as long as we're not content that we can continue to repeat. Repeatability means improvement. It doesn't mean staying the same.

Okay. As you look at the third quarter sales and maybe what you had thought fourth quarter was going to look like, but really, we're not -- we're late enough in the quarter. We're not talking about the quarter. But as you talk about maybe the next quarter or 2 going forward is what is the composition of some of that kind of deal flow look like? Are you seeing any changes around customer interest in new products? Have you had any impact, I would say, benefit from maybe less downsell coming off of some of the 0 interest rate cohorts like some of my other companies, have kind of gotten through that headwind. I don't know if that's a benefit there. Or is there anything that's kind of like different, I guess, as you think about deal composition today than maybe a year or 2 ago?

I mean Agentic is definitely a part of the deals, right, in terms of the discussion. I don't -- it's not at a point where as we've said like it's meaningfully impacting the revenue number per se. But I give this example. So in our third quarter earnings call, when the CIO for one of our top customers says, hey, with Agentic, we love the vision and with process orchestration, you're a part of our road map for the next 6 years, 7 years, right? That gives them confidence to start upselling even on what they're doing today. There are deals that are out there that are -- some deals, some of our pilots and POCs, yes, they're starting to come into our pipeline. We're excited about that, right, in terms of where that could be.

So I look at it, I don't really see core RPA deals anymore alone. I think RPA is a great entry point for the company, but I look at our growth coming from the entire breadth of our platform and including for products that we don't talk a lot about, test automation. We're super excited. If you go out and you look at the Everest Peak matrix, we are now a leader in test automation. So application testing. Like those are things that are under the radar that are very synergistic. And if you go to a CIO and you can say, wait, I can transform my processes and I can automate manual testing using contemporary software, not kind of legacy software.

They're super excited about that. So there's multiple vectors, I think, of our platform that are speaking to and manifesting itself into our pipeline. Your question around downsell and 0 interest rate environments, I think the world is super volatile. So I think it depends on the quarter in terms of composition of what you see in terms of those things. Like that is why it's super important for us to just continue to deliver and drive adoption every single day rather than worrying about cohorts of past customers.

No, I just asked the question because as I talked about the message in the last 6 months or so, 9 months maybe, your level of confidence around the sales improvement and what's happening. It's just -- it's been unique. It's been different than what we've seen in the last couple of years. So obviously great to see. I guess how much of the -- you kind of already talked about this is on the current conversation is agentic versus RPA. Do you have any of those conversations anymore? I know you're not focused on it, but is it really all agentic? Or is there a few stragglers out there?

No. I mean, agentic definitely is like the hot buzzword. So there's no -- from a marketing entrance standpoint, say agentic and a door will open, right? I think that's kind of -- that is the truth of it. The part that's hard for me to articulate to an investor, and I've used this example, and I'll try it again, is one of our customers, they wanted to automate their operational flow of financial reconciliations, okay? They didn't ask us at the end of -- we did a workshop. At the end of it, there was 100 ideas that came out. They didn't care if it was a robot or an agent. They didn't say, so how many robots did we have? How many agents did we discover? It's just how many process steps of the process can we automate.

And that happened to be 50-50. So what's interesting is like you kind of lead with transformation, you can lead with the agentic line, but it's kind of the tip of the spear that pulls the full platform through on our best engagements with customers, right? And that's really what we're seeing from a customer response. And I mean, like if you go and ask -- if you go to any CFO and you say, like agent versus robot, they're just going to say ROI. And I think our job is to show how the combination of that maximizes ROI. But agentic is definitely the thing that gets response.

Okay. So if I look back at the third quarter results, one of the verticals, I think that surprised -- certainly surprised us, even though it was in our preview because we saw the data was your federal business, had a really good quarter, especially in a segment that's seen lumpiness all year for a variety of reasons, right? General disruptions, whether it's DOGE or government shutdown, obviously, clearly saw plenty of challenges there over the last 12 months. But how has that opportunity kind of changed for you in general? Is some of those items that have been implemented the last year, that thought process, is that what maybe drove a slightly better quarter and better opportunities there? Or is there maybe a more complex answer than that?

I think it's more complex. Like -- so I want to dispel a few doubts. There was nothing that, hey, you had a slow second quarter, so we got your third quarter strength was a catch-up in federal. That's not true, to be super clear. the federal business, I think, is -- it's more complex because there are areas that I was in the Pentagon. And I actually -- as I was coming in, one of the undersecretaries like gave me an introduction to the DOGE team member that was there. And their view is for that branch of government, DOGE was kind of like, no, I'm just kind of an oversight now. Like I don't really have a budget. I don't have a team because we're like, hey, we should get together and talk through where your efficiencies. And he was pointing me back to the undersecretary, just as an example.

There are other branches of government where some of the DOGE members have become staples into that area. So the way they scrutinize the deal looks differently, right? So when you add it all together, it's kind of like a normal functioning business on average, right? Just like the customer, you end up areas where you got a really tough procurement person. You got an area where you got a super exciting initiative that is business-driven, that is kind of budgets are available and we're moving. And then you got areas where there's harder problems that you have to go to solve. So that's why we've talked about the federal as kind of a new normal. We haven't talked about it as a onetime kicker to the company. I just -- just like the overall economy, we talk about it as variable.

I consider the federal government as variable. I think we're going to have different branches in different points and different times that are there. One last plug is the Social Security administration, they had a great renewal and even additional software that goes there. Why I mentioned that is that was one of the biggest things last year around February, March with a tweet that came out that said, we're using 40% of Salesforce, 50% of UiPath, whatever it could be, don't quote those numbers. Those are metaphorical or examples. And they start digging into it. And they say, no, we're getting a ton of value for it. We got a big discount. We would rather get more usage than try to restructure the contract. So it really depends on kind of the moment and the agency.

Probably 2 more questions for me and then happy to take any questions from the audience. As we think about from a model perspective, this last year had a 200 basis point headwind from the shift to the cloud on the gross margin line. How do you see kind of cloud adoption growing with the new agentic functionality? And should we expect some similar types of headwinds around margins in the next couple of years? And I know you're not guiding to next year, just kind of...

Yes. Look, I think we're going to be doing cloud. Like cloud is going to continue to move. So I think SaaS headwinds, I think they'll continue to exist. There are no -- I feel like we've been pretty consistent from what we've talked about. Especially with agentic, you got to kind of -- like it's even a bigger drive to cloud in our mind. So I think SaaS headwinds, I think, will continue to exist. And 606 revenue recognition versus kind of full ratable, like those are all parts of the discussion depending on the technology and pricing, et cetera, that will be a part of every software company, but particularly ours as we go forward.

In terms of margin, I look at margin -- gross margins, we performed very well. We look out at the agentic and the AI era. At this moment, we don't see anything that like is something that we're super worried about. I think we're going to have to continue to monitor and we'll update our models accordingly for that. But overall operating margins, we feel good about. We feel like we can invest in engineering and sales and continue to drive productivity and everything that is not engineering and sales. And that, on average, we feel like gives us continued operating leverage that we can realize across the bottom line. Some of that is our own internal identification that we are driving, and it's super exciting in terms of the productivity that, that can yield for us as well.

Yes. Your operating margin leverage this last year, at least with current guidance in the fourth quarter is, we'll call it, 600 basis points and number, I believe, at least. And I thought that was a super interesting year because you're above 20% now. We expect it to be over 20% with the fourth quarter results. And growth looks like it's really stabilized after a couple of rough years. And natural question I get is how do you get the leverage from this level and you feel pretty confident about.

Yes. I think G&A -- so R&D, I think you're going to see us continue to invest. I think sales and marketing is too big of the line, right, to say we're going to invest in sales and marketing. We're going to invest in customer growth sales, meaning things that drive growth. Things that drive processes, we're not in -- like that's areas of pockets of efficiency. And then the last thing I would just point out is like we're super proud of is GAAP profitability. Third quarter was the first third quarter in the history of the company for GAAP profitable. We talked about kind of being on track to be GAAP profitable overall. And now we have like -- I think we are looking at capital allocation, not just on a non-GAAP basis, but inclusive of stock-based compensation, et cetera. For us, that's overall capital efficiency. And that positive GAAP profitability to me is also scalable as we continue to grow.

Fantastic. With that, we do have a few minutes left. Happy to take any questions from the audience, if there are any.

I'm curious how you are seeing the global system integrator market and some of the BPO players look to leverage UiPath technology to accelerate some of the roadmaps of their process transformation.

Yes, it's a great question. The GSI framework, I think every GSI obviously is working on their strategy, right? Our strategy has been let's double down on a few of them. I think the mistake we've done in the past is just bandwidth. It's not that we like one, we don't like one, one likes us, another one doesn't like us. It's really about bandwidth about return on our investment and theirs of where can we get like real growth that comes back to it. So the concept of saying you just got to give money across 7 GSIs, 8 GSIs, like that's just dilutive to the company versus really doubling down on a few.

If I look at Deloitte, I think Deloitte has become a really great GSI partner, an example of a great GSI partner for us. We are embedded in sales motions together. The S/4HANA upgrade is a massive initiative for them. We are deep within them to drive higher clean core percentages for customer outcomes, inclusive of including test automation in a lot of their S/4HANA migrations. There's a real 360 relationship and a core strategy that's we see ROI too, right? We're excited about other GSIs as well. I won't go through all of them for time. But I think we're going to be really particular about going deep in a few and potentially having to deprioritize others to get that.

In terms of the BPO market, same discussion. I think there are certain companies that we've developed a great relationship with. And there's real opportunity there that they see, particularly with Maestro. If you think about complex, where do you need observability? where do you need of automation? And where do you need to be able to govern agents, robots and humans? What's more complicated than a BPO shop that's doing that across hundreds or thousands of customers. And that is a big area of partnership for us and look forward to giving some announcements on that front.

Maybe one follow-up on that. On the Maestro platform, here's how your competitive landscape changing as [indiscernible] market and how much of differentiation [indiscernible] capabilities with RPA coupled with the [indiscernible] agentic platform.

So the first is I think people -- I think we contributed to this to be clear, but I think that it's important. Agentic orchestration and process orchestration are 2 different things. Agentic orchestration is focused on the agents and being able to trigger them, govern them, et cetera, independently. Process orchestration is stringing the end-to-end process together and being able to govern deterministic, probabilistic and human in that area. So when Scott was asking about human in the loop, agentic orchestration is not around human-in-the-loop capabilities, right? So in our minds, we are contemporizing kind of a -- we're leading in a process observability market, and we're contemporizing what was BPMN to a certain extent, and that gives agentic orchestration capabilities as well. That's how we think about it.

Snowflake just announced that acquisition of Observe. Is that likely to be a comparative....

No, I think yes, I think data observability is -- so and I always say this to Elise, like I should have a plaque of Maestro behind us. And I really encourage everybody. There's a fireside chat that's on our Investor Relations site from a couple of years ago -- from a couple of months ago. We can post it or we can put a little blog, Elise or I can send it back. I want you guys to just think about any process. And I said -- and think about a whiteboard. And I said, go map out the -- you would put a process map, right?

Put boxes on the thing saying, extract document, upload document here, if the document, if this matches, go here, et cetera. That's different than what Snowflake is doing. I'm saying in process map, that's what we're doing. Now imagine that, that process map is not static. It's live. And you can see the software, the robot go and grab 100 documents. You can monitor how many of that got uploaded. You can see how many is sitting in Ashim Gupta or Scott Berg's queue to review in human in the loop, right? And you can see that in the end-to-end process. That is kind of what we talk about when we're saying process observability.

Probably time for one last one.

Your questions were that good.

I guess. Well, with that, we'll leave 2 minutes for everyone to fight the elevators. Ashim, I want to thank you for the time.

Thank you so much, Scott.

So let's start -- like to get everyone on the same page, just Ashim, you had like really healthy Q3 results last week. Shares were up a lot. So everyone is happy probably in the organization, that's nice to see as well. But to get everyone on the same page, can you talk about the highlights you saw last quarter?

Yes. I think -- so our third quarter really showed the -- to me, the confluence of really strong execution and the beginnings of what Agentic, the combination of Agentic and deterministic automation, can mean in the marketplace. So revenue of $411 million, up 16%, strong operating income, strong free cash flow continuing for the company. And it was actually our first -- our first third quarter of GAAP profitability and has put us on track to be GAAP profitable for the entire year. When you look at the top line, I always go first to our large customers and our customers greater than $100,000, more than 2,500 customers now, that's growing double digits year-over-year.

And customers greater than $1 million, 330-plus customers, that's growing double digits as well year-over-year. So when you kind of step back and you look as to why, Raimo, I think very good progress around proof of concepts and POCs and pilots within Agentic. But without even a direct impact of material Agentic revenue, people are looking at our platform, the combination of deterministic automation, probabilistic or Agentic automation and process orchestration. And just the feedback has been so positive that we are a staple in people's enterprise architecture. And we'll continue to expand with them over the coming years.

And then the -- just to stay high level a little bit longer, like federal was a big discussion all year, DOGE at the beginning of the year, and there was a lot of kind of noise on -- for our industry, then the shutdown as well, like, how does it play out for you guys?

It's a dynamic environment in the public sector. I actually was just at the Pentagon on Monday. And with customer meetings, like they're going through budgets and not unlike a lot of corporations, like you go in and people are pulling their hair out in terms of challenges that they're facing and areas that are there and different dynamics. With all of that said, our teams are executing incredibly well. Like we are connected to keep people in the government under secretaries of the Department of the Defense or different functional leadership there.

We closed several key deals, the Social Security Administration, a key deal that was -- that closed there, the Coast Guard. So we're relevant across a lot of major industry -- major sectors of the government. I think it's going to continue to be dynamic. And I think it's a very good long-term opportunity for us. Our goal is to stay connected, continue to drive value and execute on the projects that we're getting today. There is a lot of momentum that we can continue to capitalize on in the coming years.

And to you like for the quarter itself, did the shutdown impact you? So I mean, in the quarter negatively, it could kind of drag into the next quarter. But like was there any impact?

No. I mean the shutdown was -- it was topics of discussion, but it wasn't anything that impacted transactions both positively or negatively for us. I'll just kind of give 2 reasons as to why. The first is a lot of our software is in mission-critical areas that are protected from areas like the shutdown that is there. And then the second is the Department of Defense is a really big customer for us, and they are also somewhat shielded from a lot of the different areas.

That being said, having walked through the floors of the Pentagon, like a lot of people were impacted, so we're very empathetic I think a lot when I say mission-critical areas, but our software was protected from that standpoint.

Yes. And then last question on this kind of more bigger picture subject is like beyond federal now, like what do you -- if you talk to customers in other regions, other industries, like how is end-demand kind of playing out at the moment? What are you hearing there?

The best thing I can say is variable. I mean like what we're seeing in our -- maybe I'll do it from 2 perspectives. What I see kind of from our -- hear from our field and what I hear directly from our customers in our customer visits. Directly from customers, they kind of have -- their budgets are really getting scrutinized for high ROI items. So you listen to kind of like there's not a lot of excess fat for experimentation and innovation. They're in the budgets. And people are preparing themselves for different outcomes of the economy because I don't think there's a consensus view about what the economic outlook looks like for next year.

At the same time, that is also spurring really renewed interest on meaningful ROI and productivity projects. So industrial customers as they're getting their cost out targets and we're very -- we're close to them. I think that presents opportunities. I think there's other customers where there's opportunity and they're saying, listen, we at the same time, may not take as large of a bite at the apple as what we've been doing. So I think it's a dynamic environment.

Then I look at our field, I look at our activity. We were really pleased in the third quarter with kind of commercial momentum in the commercial enterprises, Americas, particularly in health care and financial services. I think we did see pockets of strength in international, like Australia and New Zealand. But then there are markets that are a little bit more challenged internationally.

Yes. Okay. Perfect. And then Shifting gear a little bit, and I'm happy you are here as the CFO because I wanted to talk about AI. And I love Daniel to bit, but his explanation is almost getting too technical for me. Like how would you describe UiPath's role in this new AI now?

So a minister from really basic building blocks. AI is just unequivocally a net -- is a meaningfully net opportunity for UiPath's growth in the coming years. And I think there's a direct and an indirect basis. And I'm going to unpack it in 3 steps.

The first step is I think sometimes we forget what does UiPath do. When I ask people that they'll tell me what we did in 2017, which is RPA.

Yes.

Our -- what UiPath is an AI-powered platform that helps businesses automate and transform processes. So our goal is to drive outcomes that yields productivity for companies by emulating what humans do, to automate processes and reduce the cost of transactions and process ownership within a company.

Now dissect what is a process. A process is a series of 20 steps. Within those 20 steps, maybe 30% of them or 40% of them can be done on rule-based, task-based deterministic type steps. Another 40% or 50% of them can be done, but they need reasoning power. It's probabilistic in what needs to happen.

You could deny this claim, you may not decline this claim. It's not a rules based. If it is above this value denied. It is -- there's multiple factors that require reasoning and probabilistic waiting to go after that.

Three years ago, we automated those first 30%, and we scaled the company from next to nothing to $1 billion plus in 4 years with those steps. What is amazing is now we can go and automate those next series of steps and continue to drive that automation.

What is missed -- so there's a direct impact of AI? And what is AI? The ability to reason like a human to complete process transactions, if I keep it nontechnical right?

That's why you're here.

And then the second piece of it is, though, for UiPath, three years ago, you may have looked at this 20 step process and said, "I can automate 6 steps of the process, and it's worth it." But there are also times you'd say, "If I automate 6 steps of the process, it's not worth it. I still don't get the productivity. But if I automate 12 steps of the process or 15 steps of the process, that is really worth it." So what agenetic is also doing is it's coming back around and saying, you can automate those. So it's actually pulling back through deterministic opportunities as well. So that's the second piece.

The third piece is, AI has now created another element into the architecture of business processes that we all have to contend with. If you thought about years ago, you deal with infrastructure applications, right? And then you deal with integrations. And I would put RPA in kind of an integrated type technology that stitches together systems, et cetera. Now AI has another area. So what is becoming super important is process orchestration. And I want to differentiate between process orchestration and Agentic orchestration.

Agentic orchestration, many people do, it's governance frameworks around agents. UiPath does that as well. But process orchestration means now you have a 15 -- a 20-step process. Realistically, it's a 200-step process. You have humans, robots and agents working together, how are you going to monitor it? How are you going to govern that? How are you going to visualize it?

That is what Maestro is, a product that we GA-ed earlier this year that allows companies to do that. So AI has now unlocked a whole set of processes and opportunities for UiPath and for our customers. And what's super important is it revitalizes demand for deterministic as much as it presents an opportunity for a new revenue stream in Agentic. Does that make sense?

Yes, totally. The one question I have on that one, Agentic is where everyone wants to go. And that sounds like super cool, new, exciting -- but the one thing that I hear from the field is where customers are struggling is that Agentic, the outcomes are not as predictable as deterministic. And so in your conversation, where are customers on that journey of understanding Agentic is different and kind of having the right guard rails, et cetera, to kind of still come to outcomes that they actually kind of want?

It's the early innings, but it's the early innings with progress. So to use kind of a football metaphor or a soccer metaphor, you're not in the first -- you're not right at kickoff where players still haven't moved, but you're not near half time, right? You're not near kind of the midpoint of the game. And the reason for that, I think, is a couple of areas. One is everybody is now contending as I said, another entity into process architecture, et cetera. So there's security. There's governance. There's things that have to be unlocked with it.

The second is just understanding where do they want to deploy it. Here's what I'm excited about at UiPath. We have 900 companies that are using UiPath and building agents upon UiPath, right? 750 plus companies that are doing that. So they're in the mode of experimentation. Some of those experiments have moved to pilots. Some of those pilots have moved to production, and a few of them have also converted to orders. So in my mind, kind of the spear is moving forward with customers that have like a very progressive culture, a SaaS execution culture, they are converting, and we're seeing great examples.

Reputable health care customer. They're using our Agentic platform with Deterministic to clear 140,000 provider claims that are in backlog. That's real progress and a real goal that is coming forth. A cybersecurity firm is using both deterministic automation and what we call IXP, applying different models to the right documents, to get to the next level of productivity around something as -- a process that's been forever like procure to pay. But there are many customers who are still in the early stages, setting up their guardrails that are there.

And I think that's healthy personally. I think that sometimes when you get the euphoric adoption, it doesn't last. But I think we're in an area of sustainable kind of a really sustainable demand trend, which is super exciting for us.

And for you as the CFO, how do you think about monetization like more -- I don't want to number I'm more thinking about the vehicles of monetization in terms of is it like a SKU model? Is it a consumption model? How do you think about that?

All of the above. I think -- and I know that drives people crazy. So I'll break it down a little bit. Our pricing model has always had -- first of all, Agentic is a form of AI. So I really look at our platform right now. It's kind of the core server-based RPA, API type offering that we did between 2017 and 2021, or in 2019, we already started with AI with things like document processing. We expanded into autopilot, semantically being able to develop faster.

And we're monetizing those things through AI units and Agentic units. I think -- sorry, AI units. Now Agentic is another way that we can create a consumable in a subscription type way that's kind of a use it or lose it in terms of where it is, but then also put safeguards for COGS or for cost of goods sold on a consumption layer, if the people go crazy and want to run 1 billion agents all at once. My -- when I look at it is that is our pricing scheme. Our monetization scheme is twofold. One, lead with Agentic because it will pull forward the rest of the platform, for existing customers. For new customers, get them on to deterministic because once they automate those first 3 steps, they're going to look to the right and say, "Oh, there's an Agentic opportunity."

So that is our monetization that we can go after. And then the third piece is focus on ROI. I think the more you focus on ROI, people are less worried about how am I pricing this but more is the net price of the bill of material justified against the ROI that they're getting. And that's kind of been our philosophy so far.

And just wanted to ask, and I'm not kind of trying to get the question for my sessions that I had later today. But like if I listen to the software industry at the moment, like that message of I want to be the platform, I'm going to help you, et cetera, you're going to get from quite a few players. Like how do you fit? Is it going to be competitive? Are you all working together, but for different outcomes, like how do you think about that going forward?

With different entities and different customers?

Yes. And different software vendors as well.

Yes. I think it's a great question. So let me tell you where I think we're uniquely positioned. And then I'll tell you where I think -- how we work with others that's there. Technologically, we are uniquely positioned because we are, to me, the only platform that has RPA, API, Agentic and process orchestration, underscored with AI products like document understanding, advanced document understanding like IXP, test automation, et cetera.

So I think we have a really broad platform from there. Where I look at where we can play with -- where I look at automation, I look at 4 categories: personal productivity, in-app automation, cross-application, complex enterprise-grade automation and verticalization. Okay?

Within kind of citizen development personal productivity, we'll cooperate with anybody. We want -- that's not our bread and butter. We can do it, but we're not really kind of moving into that area cross application, same thing. We will integrate and we will create connectors with the partnership with SAP to help people automate more even within SAP that is there, but that's not a core area for us.

The third area around cross application, I think I'm going to double-click on. We want to own that area. So we want to compete. And verticalization, we want to compete. We're going to pick verticals that we are uniquely positioned to go after. So we're not going after ITSM, that's ServiceNow. But revenue cycle management, claims denials. Those are areas that we can create vertical solutions and really scale.

How do we then, just to go back to the cross application area? I think on deterministic, we are the technology. On Agentic, people want to be able to build and deploy agents on our area, and we will integrate with every model provider that is there. We created our partnership with OpenAI, created partnerships with Anthropic, with Gemini. We have partnerships in place. So models can be used by anybody.

The question then is how do you kind of integrate with the rest of the ecosystem we actually love -- we don't mind if people go and build agents on other platforms, where we want to win is in process orchestration, to orchestrate UiPath robust, UiPath agents, third-party agents and humans across. And in that area, honestly, there's -- we are very uniquely positioned to win in that area.

Yes. Okay. And then the -- yes, okay, makes sense. I'll leave it there. The other thing I wanted to discuss with you is like you guys from a more organizational perspective, have been on a journey, especially last year, there was a lot of disruption. It looks like it's getting better this year. But like maybe for the audience, it's good to understand like where you guys came from, what course did and then the changes you took to kind of try to solve it.

So Daniel came back into the CEO realm in May of last year, May, June of last year. I think when we look at it, it was very simple. I think we went through an era where we scaled so fast. I think we got an enamored -- UiPath got enamored by putting big company structure and big company processes across. And I don't think that's a wrong thing to do for the right company. We were just not the right company for that, for those processes. And so I think what happened well in that time frame is connecting and selling to the C level.

What happened incorrectly is we created too much intermediary, both internal structure and strategically between ourselves and the users and the people driving process transformation on the ground within our customers. And we lost a little bit of that innovation. So what did we do? We have really kind of gone, I wouldn't say nuclear, but we were pretty aggressive in removing central organizations.

If your hands are on a keyboard, generating code, if your hands are with -- are shaking -- or leading a customer towards transformation, we are investing heavily. Go to our website, we're hiring, we're investing. If you're in the middle of those two, there's some necessity there. But we really took a very strong restructuring mode to remove that. And that's created twofold effects. We're closer to our deals. We are closer to adoption, and we are farther ahead. We walk into a quarter, not talking about the deals this quarter, we're talking about the deals next quarter.

We're talking about the renewals 6 quarters from now. And I think that operational rigor that's been brought to the company is a reason why you see net new ARR actually stabilized, which I didn't mention in the first question. It's net positive now. And so UiPath, kind of, one of the themes where we're not the same company as 2 or 3 years ago. Not just product-wise, but execution-wise, and you can start seeing that in the metrics as they begin to inflect.

And the -- if you think about it, the -- on the one hand, you kind of removed a little bit the guys to kind of create like an empire -- sorry, if I say it like that. On the other hand, your momentum up market actually continues to grow better. That almost like -- it sounds like -- I remember when you did the changes last year, it was like we're going back to our roots, et cetera. But now the momentum is coming up, like how do you explain that?

I'm still learning in the business world around things, I would say. But the one rule that I've found is if you focus on your technology and you focus on your customer, like only good things happen. And I think that has been our premise. So in my mind, it's not a question about an empire or a person or a set of activities about it's just how much of your energy and your resources are going towards innovation and helping your customers win. And I think the more that you do there, the ROI is kind of like very natural.

So from our standpoint, that signal with the field, there's an example for a customer, big Agentic use case. And there were a couple of hundred thousand dollars of a pilot. They went and they ran their first set of COGS, it came in at like $150 million estimate. I think 3 years ago, that would have been lost in the system. One day later, Chief Product Officer engaged. Our 4 deployed engineers and our services team are moving into that customer. We realize that they can do things through batch coring and from there, those COGS have come down to a point where that POC is now considered a wild success.

And now it's a question of obviously getting it into production and getting the order, but that's just a very practical example to say when you can move up and down from your customer to your team and solve their problems, you unlock demand. Demand is there. It just needs to be unchained.

Yes. And then just as we track the progress there, like the 1 thing we all want to know is like, okay, what's the metrics, what are we looking at? Is it -- as you said, net new ARR, is that still kind of the one you look at?

Yes. Well, net new ARR to me is the primary metric. ARR is the primary metric that drives the company. I think net new ARR is a good metric and the trajectory of what that looks like. I think the second too or is the momentum we have with our large customers. Customers between $100,000 and $1 million, their net dollar expansion rate is 113%. For me, that matters.

Now why? Because when you look at kind of the lower end of the market, those are customers that we bought that we kind of -- that bought from UiPath me 4 years ago, but they don't have high propensity to buy. I'm not firing customers, but they're not going to expand at the same level.

And then when you look at customers greater than $1 million, I'm super happy with the expansion of customers in that low million dollar but we have customers that are 8 figures. Those 8-figure customers with 2- or 3-year contracts, they're not going to be expanding 20% per year, right, in that way. So I would look at that $100,000 to $1 million cohort. That's a really good sign for us. And the continued momentum to push customers into the $100,000 category and the $1 million-plus category.

And then the -- is there like the one thing and then there was always kind of difficult for you is like you have 606. There's cloud and cloud mix, it kind of would make it kind of slightly easier. But how do you think -- do you think there is something that you can solve or you -- we basically -- I'm just asking for myself actually because...

It's a fair question. I think it's on us over the next period of time, like I think refreshing the models for everybody would help. That's something that's been on our mind. But ultimately, 606 revenue is done -- it's impacted by duration and deployment and volume, right? I think volume -- I think ARR is really just impacted by volume, right? And through there, there's obviously different idiosyncrasies that exist in every metric. So my view is I look at -- I still think revenue, like revenue grew 16% this quarter, right?

I think when you look at it, it's looking to guide in low double digits is where consensus has us here for the year. ARR is at 11%. So I think you have to look at -- I think if you look at revenue on a quarter it's less relevant than looking at revenue for a year, and it's less relevant than looking at it on a trailing 12-month basis, right? So I think the more aperture you give to, if you look at revenue, you have to give it in its appropriate aperture, and it correlates very well with ARR. I think the myth around ARR and revenue, there are dynamics that change it. But when I look at it today, I think we've done a very good job when you look on a trailing 12-month basis of it correlating pretty well with the AR movements that you see.

Yes, yes. Okay. So no change, but like no change like in terms of what you do, but it's more like focus us more, ARR is the kind of the metric.

Exactly.

Okay. Last couple of minutes, profitability. If you think about it, there's a lot of stuff that we want to do around AI. The world is evolving everything quickly. At the same time, like we need to think about profitable growth on rule of 40, like where are you on that journey?

Yes. So one is I want to be unequivocal, like we are investing for growth. There is no doubt that should be there. I think the misnomer is that in order to invest for growth, you have to not be profitable, and that's not the case. You can eat if you exercise, right, so to speak. And you can get the right foods and enough of them if it's the right food and still be healthy.

So I look at it this way. I think field sales capacity, making sure we're funding our POCs and our pilots and the success of those early areas of Agentic. We are all in, and we are aggressively investing in. I think the innovation side in R&D, we're investing in. The areas, we still have areas to leverage even within G&A and sales and marketing to fund those investments where we can fund growth but continue to drive operating leverage. You can see that in our metrics.

First year of GAAP profitability, that's where we're on track for this year. I think one of the interesting things that people aren't looking at is open our stock-based compensation percentage of revenue, how it's come down and look at the trend that you've seen there. That's with the same number of people, but just being more thoughtful about how we deploy it, right?

Second is cash allocation. I think we're really very mindful about what are the top customers we want to win. It's not going to say, all markets are all equal, all at the same time. So I'm actually really pleased with growth being our priority. Profitability being about discipline. And when you look at the metrics, stabilizing net new ARR, first year of net new ARR growing year-over-year, a little bit of FX in there, but overall, that trend is there. But it's also the first quarter of GAAP profitability, and third quarter of GAAP profitability. I think those 2 things are proof that you can do both simultaneously.

Okay. Perfect. And then last question for me on the last minute, like capital allocation. What should we view...

Yes. I think UiPath is in a really good, unique position for it. So we have $1.5 billion on our balance sheet. We've already shown that we will execute responsibly for stock buybacks, we're at $800 million of a buyback. The average price of our buyback was in the low $12. So just think about that relative to a stock that is $16, $17, I think we're $18 plus today. So we're very responsible on it. It's not just a lever for the marketing point. We really want to generate shareholder returns in everything that we do.

And at the same time, we've got -- we're generating free cash flow. So we can continue to buy back stock opportunistically. We will always evaluate that. We can deploy and tuck-in M&A, which we've also done with the acquisition of Peak this year, and we can continue to look for opportunities to do so, where we can continue to have cash and have a big wallet for the right opportunity economically.

Yes. Okay. Perfect. 10 seconds, like I'm German, so I need to give some time. That was great closing statement as well. Good to see you again.

Thanks so much.

Thank you.

Thank you.

Greetings, and welcome to the UiPath Third Quarter 2026 Earnings Conference Call. [Operator Instructions] Please note that this call is being recorded.

I will now turn the conference over to your host, Allise Furlani, Vice President of Investor Relations. Thank you. You may begin.

Good afternoon, and thank you for joining us today to review UiPath's third quarter fiscal 2026 financial results which we announced in our earnings press release issued after the close of the market today. On the call with me are Daniel Dines, Founder and Chief Executive Officer; and Ashim Gupta, Chief Operating and Financial Officer, to deliver our prepared comments and answer questions.

Our earnings press release and financial supplemental materials are posted on the UiPath's Investor Relations website. These materials include GAAP to non-GAAP reconciliations. We will be discussing non-GAAP metrics on today's call. This afternoon's call includes forward-looking statements regarding our financial guidance for the fourth quarter fiscal year 2026 and our ability to drive and accelerate future growth and operational efficiency and grow our platform product offerings and market opportunity. Actual results may differ materially from those expressed in the forward-looking statements due to many factors, and therefore, investors should not place undue reliance on these statements.

For a discussion of the material risks and uncertainties that could affect our actual results, please refer to our annual report on Form 10-K for the year ended January 31, 2025, and our subsequent reports filed with the SEC. Forward-looking statements made on this call reflect our views as of today. We undertake no obligation to update them.

I would like to highlight that this webcast is being accompanied by slides. We will post the slides and a copy of our prepared comments to our Investor Relations website immediately following the conclusion of this call. In addition, please note that all comparisons are year-over-year, unless otherwise indicated.

Now I would like to hand the call over to Daniel.

Thank you, Allise. Good afternoon, everyone. Thanks for joining us. We are pleased with our performance in the quarter, which was driven by team's focus, consistent execution and progress across our strategic priorities. Our automation strategy combining the reliability of deterministic automation with the intelligence and adaptability of agentic AI continues to align with what customers want most: trusted enterprise-grade automation that delivers tangible ROI fast.

We've always believed that our approach to automation, agents and orchestration creates a durable competitive edge. This quarter's results reinforce the value of our platform and the improved execution from our teams. We beat the high end of our guidance across all metrics levering third quarter ARR of $1.782 billion, up 11%. Further reinforcing my conviction that our business is stabilizing and being driven by $59 million in net new ARR. Revenue was $411 million, an increase of 16%.

Our disciplined approach to operational efficiency continues to strengthen profitability, resulting in our first GAAP profitable third quarter while increasing non-GAAP operating income to $88 million or a 21% margin, and we are on track to be GAAP profitable for the full year 2026 for the first time.

The momentum we are seeing in the market isn't just in our results. It's in how customers and partners are engaging with UiPath. You could see that momentum in action at FUSION where thousands of customers, partners and developers joined us in Las Vegas to see how agentic AI is delivering measurable ROI today. We showcased advancements across the UiPath platform, including new integrations with partners like OpenAI, Microsoft, NVIDIA, Google and Snowflake and real customer storage where we had leading enterprises across key verticals, sharing how they are transforming mission-critical processes with Agentic Automation.

Customers tell us they get the most impact from a unified platform, bringing together deterministic automation, agentic intelligence and Process Orchestration. And we are seeing that play out that enterprises move quickly from pilots to production. Over 950 companies are developing agents, and there have been more than 365,000 processes orchestrated with Maestro across our platform.

A powerful example is one of the world's largest investment management firms which shows UiPath for Maestro's vendor-agnostic architecture and ability to connect systems. They've already demonstrated measurable impact through multiple agentic POCs integrating with ServiceNow, Confluence and specialized LLMs to orchestrate end-to-end workflows and delivering a 95% reduction in time to value and tens of millions in projected savings.

With more than [ 260 ] automations, they expanded this quarter to increase their adoption of Agentic Automation. And together with Ashling, has identified over 40 high-value use cases expected to generate more than $200 million in savings over the next 3 years.

These stories demonstrate how our innovation is meeting customers where they are and helping them scale. And it's that customer energy that inspire many of the new capabilities we announced at FUSION. One of the most exciting is UiPath ScreenPlay, where we are combining traditional RPA with the power of LLMs build more reliable automation. ScreenPlay, understands intent, build multistep plans and execute them autonomously, giving developers and business users a faster way to automate complex UI tasks.

Our ScreenPlay technology is one of the best position in the computer use benchmark World OS (sic) [ OSWorld ]. We are also helping customers move from pilot to production faster through continued enhancement to Agent Builder, including a new visual canvas for debugging and optimization and reusable templates that make automation easier to scale, that focus on usability is driving real customer success.

A great example is USI Insurance Services which selected UiPath because of our multi-agent orchestration capabilities. Working with Lydonia, they are automating a complex workflow where UiPath agents and robots process incoming requests and generate output, all orchestrated by Maestro. USI expects over $32 million in savings over the next 3 years.

As more customers scale their Agentic Automation programs, we're expanding that what agents can do, especially in document-heavy processes. This quarter, we introduced agentic capabilities to our Intelligent Extraction and Processing, our IXP product which delivers specialized extraction and validation agents that handle complex nondeterministic scenarios and reduce manual review. And with UiPath's autopilot for IXP, customers can automatically generate document template saving hours of setup time per project.

A great example is Corewell Health, which plans to leverage IXP to automate the processing of referral information into Epic, in addition to improving efficiency and accuracy, they are on track to redirect $1.5 million of labor savings this year and expect over $3 million next year.

Something like this show how our innovation is helping customers turn manual document-heavy work into intelligent automating processes. And these strong results continue to earn us third-party recognition. We were named a leader in the inaugural Gartner Magic Quadrant for intelligent document processing which we believe highlights our capabilities for data and information extraction to help enterprises unlock value from their documents in the age of AI and Agentic Automation.

We are also recently recognized as a leader in the Gartner Magic Quadrant for AI augmented software testing tools, which, in our view, validates our vision for Agentic Testing and the results our customers achieved with UiPath Test cloud. One example is [ Energie Energy ], which adopted UiPath test cloud to address testing challenges across SAP and its digital apps. With Agentic Testing and autonomous self-healing, they expect 30% better coverage, 1.5x faster cycles and almost $2.9 million in savings over 3 years.

And lastly, we are proud to be recognized by Everest Group as both a leader and star performer in the 2025 Intelligent Process Automation Platform and full suite IPAP peak metrics assessment, this recognition underscores the strength of our end-to-end platform, our innovation in Agentic Automation and AI integration and the tangible business impact we are delivering for customers.

These recognitions highlight the strength of our unified platform, which connects every layer of automation from discovery to orchestration within a single governed system. They also reinforce what we hear directly from customers. The power of our platform comes from how it works together.

And at the center of that platform is Maestro, our orchestration engine. Maestro builds beyond managing automation it serves as the control plane for how work is orchestrated across the enterprise, powering through end-to-end automation at scale. A powerful example is the leading U.S. managed care provider that is leveraging UiPath agents, robots and Maestro to tackle a backlog of more than 140,000 provided appeals. They automated the entire workflow using agents to classify forms, robots to handle processing and Maestro to orchestrate the process. The result is a streamlined operation targeting 80% autonomy in year 1.

As Maestro becomes more deeply embedded in customer operations, we are continuing to enhance its capabilities with new features like case management and process apps, Case management helps customers model and manage now-running processes, while process apps enable customers to build tailored end user experiences with real-time visibility and actionable insights to drive proactive operational improvement.

In order to realize the value of agentic, customers need a strong foundation in deterministic. And this quarter, we announced the general availability of API workflows, which delivers API-centric automations that complement RPA and agents. We believe this strengthens our position as one of the industry's most comprehensive automation platform. We are continuing to expand our cloud footprint in important markets like Switzerland.

And at GITEX, we announced the launch of Automation Cloud in the UAE. This expansion gives customers the ability to run automation, AI agents and orchestration in a secure, locally hosted environment that meets regional data residency and governance requirement. The power of our platform really comes to life when it's applied to industry-specific challenges. We are focused on building vertical solutions that help customers accelerate our constant ROI.

The capabilities we gained through our Peak acquisition earlier this year, are extending these vertical capabilities. By combining their industry-leading pricing and inventory optimization technology with Maestro and our broader platform capabilities with creating an agentic merchandising, pricing and inventory solution with Debenhams Group. We are taking this joint agentic solution to market with other leading retailers and manufacturers expanding the reach of our platform across key verticals.

These industry solutions are just one way we're helping customers realize faster value. We're also enabling them through deep collaborations with global technology leaders. At FUSION, this collaboration came to life through new integrations, including with Microsoft Azure AI Foundry, enabling UiPath agents to work seamlessly with Azure agents and models. Using model context protocol, UiPath, expands integrations with Microsoft Copilot and empowers organizations with the trust and governance needed to run agents at scale. We announced a collaboration with OpenAI to deliver a ChatGPT connector bringing OpenAI's Frontier Model directly into enterprise workflows, simplifying AI agent development and accelerating time to value.

We also launched a new conversational agent with Google's Gemini models enabling natural voice-driven automation without complex coding. Additionally, we introduced a new integration with NVIDIA, allowing organizations to enhance high trust workflows like fraud detection and health care with AI models deployed through NVIDIA NIM Microservices. And finally, we partner with Snowflake to combine Agentic Automation with Snowflake Cortex AI, helping businesses to data insights into fast autonomous action.

While our technology partnerships expand what's possible with our platform, our go-to-market partners make that innovation real for customers. During the quarter, we expanded our collaboration with Deloitte to help organizations accelerate how they build, test and release software. By combining the genetic testing capabilities of UiPath Test Cloud with Deloitte Ascend, we are transforming the testing life cycle with Agentic AI to automate repetitive tasks, we take changes and execute tests autonomously. And with autopilot for testers and Agent Builder embedded in Ascend, Deloitte teams can leverage over 1,500 prebuilt testing bots and AI agent.

Lastly, we are encouraged by our federal sector performance this quarter, where efficiency mandates are creating a long-term tailwind for automation. Highlights this quarter include expansions with the U.S. Coast Guard to modernize core systems and improve mission readiness through automation and AI. The Department of Veterans Affairs which is automating disability claims and enhancing contact center service for veterans and the social security administration which is migrated to the cloud to expand to our IDP solutions to help accelerate benefits processing.

Our unified platform and innovation continue to strengthen these partnerships and underscore the significant opportunity ahead in the public sector, even as the federal purchasing environment remains dynamic with pockets of strength.

Looking ahead, our continued innovation is expanding what's possible for customers and delivering measurable results for the power of deterministic and Agentic Automation. What continues to set UiPath apart is our unified end-to-end platform architecture, delivering one connected experience from discovery to deployment with Maestro orchestrating work across systems and our built-in governance capabilities, ensuring control, compliance and trust, I am pleased with the progress we are making in improving the execution of our teams and the pace of innovation across product organization is delivering for customers. We feel well positioned as we close out the year and continue executing on our vision for the future.

With that, I'll turn the call over to Ashim.

Thank you, Daniel, and good afternoon, everyone. Before turning to the financials, I'd like to provide a quick operational update. This quarter reflects the meaningful progress we've made in sharpening execution across the strategic priorities Daniel outlined earlier this year, including strengthening customer relationships, accelerating innovation and driving operational rigor across the organization.

Through these areas of strategic focus, we have improved performance of our sales team. Deepened areas of strategic focus, we have improved performance of our sales team, deepen engagement and strengthened alignment with our customers' priorities. We're partnering earlier co-developing solutions and scaling automation faster. Our broad installed base gives us unique visibility into enterprise workflows helping customers connect people, robots and AI agents to deliver measurable outcomes at scale.

Our pace of innovation continues to accelerate, supported by deeper ecosystems, integrations and advancements in our Agentic Automation platform. Combined with disciplined execution, these efforts contributed to another solid -- another quarter of solid top line and bottom line performance, including our first GAAP profitable third quarter and putting us on track for our first GAAP profitable year.

And we're seeing customers lean in a great proof point of our end-to-end platform is a leading cybersecurity company that expanded to our agentic products. With support from our Forward Deployed Engineers, they are leveraging UiPath agents, robots, IXP and Maestro to create seamless end-to-end workflow. IXP extracts data from purchase orders across 700 vendors, robots retrieve quote details from SAP, and agent supply confidence scores before passing the data to sales order creation. Maestro orchestrates the process, ensuring speed, accuracy and control which is expected to help them improve their accuracy rate from 50% to 90%.

Turning to the quarter. Unless otherwise indicated, I will be discussing results on a non-GAAP basis, and all growth rates are year-over-year. I also want to note that since we price and sell in local currency, fluctuations in FX rates impact results. Rates have remained largely stable since the time of our last earnings call through the end of the quarter. And as a result, there is no incremental FX impact to our third quarter results.

Third quarter revenue grew to $411 million, an increase of 16%. Normalizing for the year-over-year FX tailwind of approximately $5 million, revenue grew 14%. ARR totaled $1.782 billion, an increase of 11%. This included a $6 million year-over-year FX tailwind. Net new ARR was $59 million.

We ended the quarter with approximately 10,860 customers. We continue to be successful in signing new enterprise logos that align with our strategy of targeting long-term customers with a propensity to invest, including new logos like Nuffield Health, Resurs Holding and an Australian brick manufacturer, which selected UiPath due to the breadth of our Agentic Automation platform, and they will be leveraging UiPath robots, agents, IXP, Maestro and process mining to automate sales order processing.

As with prior quarters, the vast majority of customer attrition continues to be on the lower end. Customers with $100,000 or more in ARR increased to 2,506 and continue to have a strong dollar-based net retention rates. While customers with $1 million or more in ARR increased to 333. Dollar-based gross retention remained best-in-class at 98%, and our dollar-based net retention rate was 107%, underscoring the durability of our customer base as they embrace our Agentic Automation solutions. Adjusting for FX, dollar-based net retention rate was 107%.

Remaining performance obligations increased to $1.265 billion, up 12%. Normalizing for the FX tailwind, which was approximately $20 million, RPO grew 10%. Current RPO increased to $840 million, up 17%.

Turning to expenses. We delivered third quarter overall gross margin of 85%, and software gross margin was 91%. Third quarter operating expenses were $261 million. We delivered our first GAAP profitable third quarter, with GAAP operating income of $13 million, up from the prior year GAAP operating loss of $43 million.

GAAP operating income included $71 million of stock-based compensation expense. Third quarter non-GAAP operating income was $88 million, representing a 21% margin, up more than 700 basis points year-over-year and driven by our continued focus on operational efficiency.

Third quarter non-GAAP net income was $85 million, which excludes a nonrecurring noncash tax benefit of $184 million from the release of a valuation allowance on certain deferred tax assets. Third quarter non-GAAP adjusted free cash flow was $28 million. We ended the quarter with a healthy balance sheet of $1.5 billion in cash, cash equivalents and marketable securities and no debt. Now turning to guidance.

We are pleased with the team's execution in what continues to be a variable macroeconomic environment. We continue to maintain a prudent outlook and guide to what we see in front of us. As Daniel mentioned, we are pleased with the progress of our public sector team. As a reminder, while we are encouraged by early traction with our agentic capabilities, adoption is still in its early phases, and we don't expect a material top line contribution in fiscal 2026. Lastly, since the end of the quarter, the Japanese yen has depreciated against the U.S. dollar, creating a headwind to fourth quarter guidance.

Turning to the specifics of our guide. Despite the incremental FX headwind from the yen, we are raising guidance for the progress we've made on our operating priorities and the strength we are seeing in the business. For the fourth fiscal quarter 2026, we expect revenue in the range of $462 million to $467 million. This range reflects an approximately $3 million headwind driven by FX rate movements since we provided guidance on our second quarter earnings call. ARR in the range of $1.844 billion to $1.849 billion. This range reflects an approximately $3 million headwind driven by FX rate movements since we provided guidance on our second quarter earnings call.

Non-GAAP operating income of approximately $140 million. And we expect fourth quarter basic share count to be approximately 536 million shares. And finally, we continue to expect fiscal year 2026 non-GAAP adjusted free cash flow of approximately $370 million and non-GAAP gross margin of approximately 85%. Thank you for joining us today, and we look forward to speaking with many of you during the quarter.

With that, I will now turn the call over to the operator. Operator, please poll for questions.

[Operator Instructions] Your first question comes from Bryan Bergin with TD Cowen.

Maybe just starting off here on some of the agentic solution traction. Daniel, I heard you mentioned, I think, 950-plus clients. Just first, is that comparable to the, I think, the 450 or so last quarter using Agent Builder? Or is that a broader kind of view across your agentic solutions? Just trying to get a sense of kind of that quantitative traction, if you could share that. And for the cases where you are seeing scaling past the proof of concepts, is it more about the underlying clients and their capabilities or more so about the specific types of use cases that they're pursuing?

Thanks, Bryan. Yes, we are seeing really good momentum across our agentic offering. And this creates a pull-through across the entirety of our platform. One -- we are seeing some kind of consistent buying patterns emerging from POC to pilots and to some use cases into production. I would say that it's more the highest ROI use cases are very customer specific. I don't see necessarily a single one across multiple industries or different departments. But overall, it's pretty encouraging to see the movement from -- again, from pilots into production for some of them.

Okay. And then my follow-up is on the federal business. So you had positive commentary here, probably a surprise for some given the shutdown. Just curious, was there any shutdown impact just worth calling out here in October and into November?

No, there is no direct impact from the shutdown. You got to remember, Bryan, a lot of our projects are just funded through the bills. So -- and many of them are considered -- are in critical operations like in areas like the Department of Defense, et cetera. So we had no major impact from the shutdown.

Your next question comes from Jake Roberge with William Blair.

Nice quarter. Congrats on the results. I know there's some FX impact but your Q4 guide implies that net new ARR could start growing again on a constant currency basis. Can you talk about the driver of that return to growth and just how we should think about the sustainability of net new ARR growth moving forward?

Yes. Look, I think the entire business is positive. We are really pleased with how our team executes. We see consistent execution across the board. I would like to nominate especially our teams in Americas, where we are really seeing signs of great traction, especially in the agentic. And yes, I would say it's overall, it's -- things are improving and stabilizing.

Yes. I would just echo what Daniel said as well. I think there is no magic to it. We talked about the improved execution. The focus -- the launch of the new products, we think, is going to continue to help us both, definitely indirectly in pulling through our platform and increasing stickiness and getting us deeper in customer conversations. But then as we go through time more directly. And we look at it kind of just as the business stabilizes, there's a lot of good news, both the consistency of the leadership, the talent that has been brought in as well. So those are factors that contributed. There's not one single magic button or one silver bullet that we're counting on for that.

Okay. That's helpful. And then I know you're not expecting a material contribution from AI solutions this year. But for customers like that cybersecurity company that you called out in the script that are starting to put these agents and Maestro processes into production. Can you help us understand what type of pricing uplift or monetization that you're seeing once we actually get these go-lives in production?

Yes. It's not about the pricing uplift. I think the first thing to note is we talked about the cybersecurity, Jake, is it's pulling through the entire platform. IXP, additional robots, Process Orchestration. So I think that is an important part of this. It's not agentic in isolation. It is agentic paired with the rest of our platform that really is driving value for our customers. So I see this kind of the monetization, not as a pricing uplift, but increasing stickiness, giving more conviction to deepen our platform into the architecture of our customers, as customers really like the road map and are starting to experiment with agentic and find tangible ROIs through the full breadth of our platform.

Your next question comes from Mike Turrin with Wells Fargo.

This is Austin Williams on for Michael Turrin. I just wanted to go back to U.S. Federal. I'm curious how results in 3Q came in versus your expectations at the beginning of the year related to DOGE? And then maybe just as a follow-up. Anything you can add on the OpenAI collaboration, what exactly that could drive for UiPath?

I would say that the federal business continues to be a dynamic environment for us with pockets of strength. We are really encouraged by the progress in 3Q, I think it's returning to a new normal. The deals that we mentioned in the scripts are really solid, team is executing well with focus on efficiency positioning as well. The projects are long term and strategic, not short term. We continue to be prudent in our guidance estimations about the sector.

On the OpenAI, yes, we use GPT5 across the board in our platform. We highlighted especially the use in one of the most innovative parts of our platform, the product that I mentioned in the script called ScreenPlay, which is basically our own version of computer use or operator. But the key thing here is that we can combine the reliability of UI Automation with the power of adaptability of LLM driven computer use. And I think to my knowledge, we are the only company that can succeed delivering autonomous UiPath using LLMs, using this approach.

And your next question comes from Matthew Hedberg with RBC Capital Markets.

This is Mike Richards on for Matt. Kind of building on that last question, there is a ton of excitement around all the partnerships you guys announced, and I think it validates your positioning and orchestration. So I was wondering, even beyond OpenAI, if you could just give some more details around the partnership just in terms of, is there a joint go-to-market element to any of them? I know it's early, but have you seen any pipeline build as a result of these partnerships? Just any more details so a lot of excitement around it.

I would say that at this point, our -- the partnership that we announced at FUSION are clearly into the technology-enabling partnerships. And they were driven largely by our customer needs. We always -- we praise ourselves for having an open platform that can really be flexible and customizable to customer needs. So we believe so, if we look at kind of our partnership, we believe that the foundation of delivering reliable in enterprise have different layers. So it has the data layer, ontology layer, so we -- this is where our partnership with Snowflake is shining.

We offer ourselves the automation, the RPA and API and agentic layer. But of course, with Open AI and Google we use the Frontier LLMs in order to power the agentic. We use NVIDIA for security and governance in regulated industry. So I think our goal is to create a set of partnership that offer a very solid foundation to deliver reliable AI into a secure and governed manner.

Super helpful. And then if I could just do a follow-up. I think before we talked about in the early days of the orchestration opportunity for you guys, it's been mostly agents created on UiPath. I was just wondering, have you seen more of a shift to third-party agents yet? Or is it still -- you guys are mostly orchestrating agents built within UiPath?

We are seeing many customers interested in building coded agents where we have partnered with companies like LangChain, CrewAI and LlamaIndex. And we are seeing right now a mix of agents that are hosted and managed by our platform that are both low code and coded agents. I think at this point, it's kind of too early to see us managing external agents that are built completely outside of our platform.

Your next question comes from Sanjit Singh with Morgan Stanley.

This is [ Ryan Lance ] on for Sanjit. Just with regards to the channel, you mentioned the expansion of your partnership with Deloitte. So can you just provide some additional details around how much incremental pipeline is now partner sourced relative to a year ago? And maybe just how these partnerships can help drive additional kind of AI-related product deployments next year?

Yes. One is, I think the quantum has definitely increased, but more than the quantity, it's the quality. So when you look at the S/4HANA migrations that are happening and partners like Deloitte and their presence within many of these customers, we're now involved in those conversations. And I think that, that is helping pull us through and being in the conversation about larger scale transformation processes. I think that is -- Deloitte has obviously done a really exceptional job and that partnership has been meaningful for us. But I think that's a motion that we're also seeing across our teams with various partners that are there. So the quality for me is much higher quality pipeline than just a superficial quantum or a quantity number that I would tell you about.

Okay. Great. And then just one follow-up. You guys have driven some pretty meaningful OpEx leverage throughout this year. And I'm just curious if you could provide any additional details around how you're thinking about OpEx investment next year to kind of support this AI product rollout and just broader monetization path.

Yes. Of course, I'm not going to provide anything specific regarding next year at this time. But I think putting this year in context, we'll give you a sense of the strategy that Daniel and the leadership team here are employing. So I think the first thing is we've got an OpEx leverage by not austerity but by discipline and really being super focused on where we're prioritizing. So we are actually hiring in our engineering segments. We are expanding sales capacity. As we talked about earlier this year, and last year, really, our focus has been continuing to drive efficiency across our processes, being selective about overlay functions in terms of where we're investing.

And I think that area allows us to get operating leverage while still being able to invest in key areas. So as we're looking at our platform, whether it's Forwards Deployed Engineers, whether it is more hands and legs at our customer sites or just core engineering capabilities. Those are areas that are in our investment zone. When you look at our line item, we're getting leverage across every area because we're really going to all of the cost structures that exist outside of those 3 and really seeing what are the areas of efficiency and prioritization. And that has not just given us OpEx leverage, it's also enhanced our focus and has contributed to us being better in our execution cadence as well.

And your next question comes from Scott Berg with Needham & Company.

Nice quarter. When I was at the conference a couple of months ago, some of the partners I spoke with really talked about a high level of pilots and proof of concepts that your customers are going through right now with the different types of AI functionality? And I know it's not a big part of the expectations around bookings for this year but I guess, as you've seen some of them convert to actual sales or production, are there any key drivers or levers that you've learned through some of these that you can help maybe use some of these other deal cycles you're currently going through?

Yes. We see these patterns emerging. It's -- I would say that the landscape is extremely scattered right now. We have deployed our teams in various use cases across various industries. We see some particular partners emerging in health care like, for instance, in revenue cycle management, prior authorization, claims management, in financial services, financial crimes, it's -- but again, I think it's a bit too early to mention one particular use case where we see like great replicable potential.

Understood. And then, Ashim, as you look at the fourth quarter guidance, someone else had already mentioned that the implied ARR number suggests that your net new ARR is up again on a year-over-year basis, help us kind of think through maybe the construction of that. Is that just purely a result of the improved execution that you all have been working on this year? Or is there some aspect of maybe deals that were maybe slipped from Q2 and Q3 that kind of moved in to the expectation? And I know that AI, some of the functionality there is not a big driver this year. But are there some expectations around maybe some bookings improvements in that category that's helping kind of drive what your initial guidance here is?

Yes. So let me be super clear on it. There's nothing in terms of like slip deals from third quarter or anything that is timing oriented, Scott, in any material or significant way. There's always normal course deals that move back and forth, right, deals that we're in your fourth quarter pipeline that closed early in deals that closely. That's just normal course and nothing out of the usual that I would talk about.

When you look at the year-over-year growth implied within our guidance, I think there's 3 main factors for me. The first is execution. We are seeing improved execution across the board from our sales team. We -- it is supported by good customer activity, maybe not 1 or 2 deals, but really broad-based activities of POCs and pilots and renewals that are coming in with more conviction about our long-term place in the architecture that leads to natural upsells, et cetera. There is macroeconomic environment, as I talked about in the second quarter earnings call around foreign exchange, that will play a role in that year-over-year impact that is there.

And then the third one is I just think momentum. I think that we've had a good stable base now in terms of if you look at our sales stability 6 months, 9 months after kind of the restructuring that we had completed at the beginning of the year, as well as just a little bit of the normalization around areas like the public sector, which we've said is kind of back to a new normal. No catch up, but at least at this time, a new normal.

We feel like we're still prudent on our overall assumptions on the macroeconomic environment. But it's a confluence of things that to me are at the opposite end of when net new ARR was declining, poor execution, foreign exchange having a headwind and frankly, just losing -- having too much inconsistency in our strategy as well as our organizational structure. So I'd say all 3 are contributing to that, to the stabilization of net new ARR you're seeing.

Your next question comes from Alex Zukin with Wolfe Research.

This is Arsenije on for Alex Zukin. And I guess congrats on the results and also just kind of wanted to expand on what the downtick in NRR was? Was it just weakness at the low end? Or was there anything else? And then you kind of unpack what's driving that stronger new business, but is there anything that you can kind of give us in terms of Q4 applications on that new business strength given the 3 factors you just talked about?

Yes. So look, I think there is -- I think when you look at it, definitely at the lower end of the segment. We've talked about that. It has a little bit of pressure on the net new ARR. To give you a supplemental metric, our net dollar expansion rate for customers between $100,000 and $1 million was 113%. So that can quantifiably show kind of the lower cohort having more pressure. And I think you get a little bit of the law of large numbers as well. But as we stabilize net new ARR, obviously, the rest of the metrics begin to stabilize as well. And you can see that pattern starting both with our third quarter results and, of course, kind of the guidance that we've provided here.

Got it. And then just to kind of follow up on kind of the same thing. You doubled the number of customers developing agents quarter-over-quarter and Maestro process instances. What's the uplift you've seen in those customers? And is it kind of fair to assume that into next year as that revenue ramps, it can help sustain that NRR and offset some of that low-end weakness that we've seen?

Yes. We will continue to contend, but we're in the early innings of agentic. I think what the momentum around customer activity, it has both the direct and indirect. I think the indirect definitely has started to help us and will continue to help us. customer pulling through other parts of our platform, investing in us as a longer-term part of their enterprise architecture.

In terms of direct scalable, direct monetization place from agentic. We're -- at this we'll update everybody as we kind of get into next year around assumptions there. But for the immediate short term, there's nothing material as we cited in the script.

Your next question comes from Kingsley Crane with Canaccord Genuity.

I think it's interesting if we think about the Code Red moments over the past couple of weeks, it's clear that competition among model providers is healthy, of course, that makes integration, orchestration even more valuable. Have you seen a shift in perception in the past quarter? And then from an integration activity perspective, how is heterogeneity trending?

I don't think we necessarily see yet a shift. I think you aim what -- Google Gemini release. We were using Gemini in our IXP business for quite a few quarters somehow. I think we -- as a platform, we continue to assess all the frontier LLMs, and we use, frankly, a mixture of them. And for instance, in our IXP business, we use GPT5 to understand the better the structure and intent documents, while we can use Gemini for specific extraction like multiple tables, indicated tables. This is just an example. But across the platform, we always monitor and use the best of breed LLMs.

Great. And then just a quick follow-up, Ashim, you mentioned in your prepared remarks that you're partnering earlier, your co-developing solutions earlier. That's been a big focus for this year. Just to what extent did this directly translate into improved results in Q3? And then do you think that this could be more of a future predictor of success next year?

I think a leading indicator, yes. I think any time you're closer to customers, you are innovating with them, you're solving problems. I think it helps. And I would say what our conviction is, is that ROI is going to be where choices are made. And so as we're co-innovating, as we're teaming up with partners around bigger problems, I think we have a chance to have meaningful impact around ROIs, which we feel we'll continue to do that.

I think it also shows the relevance of UiPath. If I can -- like in terms of the validation of the agentic framework, it's not marchitecture, it's real product that has real impact that both partners and customers can innovate around. And I think that is super important and a really great validation point for our product team, our sales team and all of our customer teams combined.

To answer your question around third quarter, as we've cited in the script as well in the prepared remarks, there's no meaningful impact from agentic in our near-term results. We feel like this is continued disciplined execution. What agentic continues to do is the closer you are to customers the more you are a part of their long-term road map and architecture, the more that they're willing to invest and pull through the existing parts of your platform today, and have higher level impacts within their org and that indirect impact we're definitely feeling the momentum and we're excited about it.

Your next question comes from Kirk Materne with Evercore.

This is Chirag on for Kirk. Congratulations on the strong quarter in traction. Daniel, you've talked briefly about prebuilt Agent solutions, highlighted strong modernization use cases within certain categories in the past, which verticals do you see adopting these prebuilt agentic solutions the fastest? And are you moving more aggressively into industry-specific packaged automation at all?

Yes. Our verticalized approach is getting a lot of traction within UiPath. We put a lot of focus and effort. We actually did recently rework of our product and engineering teams to better address the creation of vertical solutions. In terms of the industries, we focus on health care, financial -- and financial services primarily with again, with an accent on revenue cycle management in health care and in financial services, I would say, financial crimes, can be one of the -- know your customers, anti-money laundering type of risk cases we are seeing the most interest from our customers.

Okay. Maybe just one more. What are your thoughts on the balance between deterministic automation and agentic or LLM-based automation? And do you see this balance shifting materially at any point over the next few years or few quarters?

Look, our thesis is that they are very complementary, and they address different steps in the business process. And in many processes as long as task or workflow is well defined. People would use a deterministic automation. There is zero need to use an LLM-based. You use an LLM in order to create a deterministic automation was or to maintain it to make it -- to improve it over time, but you don't use an LLM to drive it. But of course, there are so many areas where you cannot have -- rules are either too complicated or process is too complex, you sift through tons of documents where or it involves conversational aspect of the process.

So in all these areas, LLMs are an amazing complement to the deterministic automation. And I also want to mention that the orchestration piece that combined the AI-based and deterministic with humans in the loop is an essential piece that is required in order to deliver a solution that is secure, governed for the enterprise. We continue -- we have continued saying that orchestration is the key. We announced our effort a year ago. And I'm pleased to see that across the industry, people are talking right now and realize that orchestration is a key technology in order to deliver a reliable AI.

Congrats on this quarter.

Your next question comes from Terry Tillman with Truist Securities.

This is Dominique on for Terry. So it was mentioned previously that one of the biggest customer hurdles with agentic consumption pricing is spend predictability. So have early deployments given you enough usage patterns to help customers forecast more reliably. And also just curious as to what else you all are doing to help customers get over that hurdle?

Look, we are constantly evaluating how our customers are adopting AI. And we aim to have a pricing model that really reflects the adoption of -- consumption of AI. We constantly monitor industry trends. And I would say the entire industry is dynamic at this point and is trying to figure out what's the best pricing. We are pretty flexible in our approach. We can price by components. We also are pretty flexible on understanding an outcome-based pricing can be for our customers. But again, I think we are in the early innings of really understanding consumption patterns across agentic AI at scale, I mean.

Got it. And then just as a follow-up, has the typical cycle time frame from an agentic PLC to a production deployment shortened versus earlier in the year? If so, could you just highlight what specific efforts are driving that compression? Or what do you all plan to do to accelerate it?

Yes. I think we, as a company, understand a bit better how various use cases. We are building internally solutions accelerators that can help us replicate experience across industries and verticals. And my estimation is this trend will continue to accelerate within next year. I strongly believe that the key to unlock huge scale AI consumption in enterprise is the prepackaged solutions where will -- we can meaningfully accelerate the time to value. This is why, as I mentioned before, we put a lot of emphasis in building these solutions.

Thank you. And there are no further questions at this time. So I'll now hand it back to management for closing remarks.

Thank you so much for all the questions. I would like to wish you happy holidays. And as usual, we like to hear from as many as you throughout the quarter. Thank you.

Thanks. This concludes today's call. All parties may disconnect.

Greetings, and welcome to the UiPath Second Quarter 2026 Earnings Conference Call. [Operator Instructions] Please note, this conference is being recorded. I will now turn the conference over to your host, Allise Furlani, Vice President of Investor Relations. Thank you. You may begin.

Good afternoon, and thank you for joining us today to review UiPath's second quarter fiscal 2026 financial results, which we announced in our earnings press release issued after the close of the market today. On the call with me are Daniel Dines, Founder and Chief Executive Officer; and Ashim Gupta, Chief Operating and Financial Officer, to deliver our prepared comments and answer questions.

Our earnings release and financial supplemental materials are posted on the UiPath Investor Relations website. These materials include GAAP to non-GAAP reconciliations. We will be discussing non-GAAP metrics on today's call. This afternoon's call includes forward-looking statements regarding our financial guidance for the third quarter and full fiscal year 2026 and our ability to drive and accelerate future growth and operational efficiency and grow our platform, product offerings and market opportunity. Actual results may differ materially from those expressed in the forward-looking statements due to many factors, and therefore, investors should not place undue reliance on these statements.

For a discussion of the material risks and uncertainties that could affect our actual results, please refer to our annual report on Form 10-K for the year ended January 31, 2025, and our subsequent reports filed with the SEC. Forward-looking statements made on this call reflect our views as of today. We undertake no obligation to update them.

I would like to highlight that this webcast is being accompanied by slides. We will post the slides and a copy of our prepared comments to our Investor Relations website immediately following the conclusion of this call. In addition, please note that all comparisons are year-over-year unless otherwise indicated. Now I would like to hand the call over to Daniel.

Thank you, Allise. Good afternoon, everyone. Thanks for joining us. I want to begin by thanking our partners and customers for their continued trust. I also want to thank the entire UiPath team for their progress on improving execution, consistency and discipline. And the progress is evident in our platform's innovation, customer outcomes and financial results. Together, we are shaping the future of enterprise transformation.

In every conversation with customers and partners, the message is clear. Automation and AI are stronger together. Our deterministic foundation, enterprise-grade RPA and API automation capabilities deliver the trust, scale and reliability mission-critical processes demand. On top of that, our leading AI capabilities, Intelligent Document Processing, or IDP, and our Agentic AI offerings that can reason, plan and act bring adaptability, intelligence and speed.

In addition, what brings it all together is orchestration. With UiPath Maestro, we unify agents, robots and people across systems with governance and transparency so outcomes are measurable and repeatable. This combination is delivering tangible ROI, fueling increasing commercial momentum and positioning UiPath to lead in this new era of automation.

It has also translated into strong second quarter results, where we exceeded the high end of our guidance across all key financial metrics. Second quarter ARR grew 11% to $1.723 billion, driven by $31 million in net new ARR. Second quarter revenue was $362 million, an increase of 14% from the prior year period. Importantly, our AI and Agentic solutions are helping us win deals and increase deal sizes faster than traditional automation engagements and now represent a growing share of commercial activity.

We continue to deliver growth while driving operational efficiencies across the organization. Non-GAAP operating income increased to $62 million, representing a 17% margin and an improvement of more than 1,500 basis points year-over-year, reflecting the operating leverage and discipline we are achieving in the business while capturing the benefits of agentifying UiPath from within.

This momentum reflects the work we've completed to elevate execution. Over the past year, we rebuilt our go-to-market for scale, stabilizing our structure, adding sales capacity and specialists and instituting value-based plays with tighter pipeline inspection. We are seeing the impact: higher-quality pipeline, more predictable forecasting, better customer adoption and implementation and larger multi-solution opportunities, especially where Agentic and IDP land on top of our RPA and API foundation.

You can see this execution showing up in the field. We continue to acquire high-quality new logos. For new customers, deterministic automation remains a high ROI, low barrier entry point. That's why over 95% of new logos this quarter included our core automation capabilities. For our large installed base, our new AI and agentic capabilities are translating into real momentum. Since launching our Agentic platform, customers executed almost 1 million agent runs. Maestro orchestrated over 170,000 process instances, and over 450 customers are actively developing agents.

Now let me share a few examples of how customers are using these capabilities in practice. One example is Voya Financial. After automating more than 100 processes, they expanded this quarter to adopt our Agentic capabilities. With Agent Builder, they are targeting over 40 high-impact use cases in accidental claims, and Maestro is orchestrating the workflows.

Another great example is an international mobility service provider, which shows how quickly customers can scale Agentic Automation across their business. This quarter, they purchased our Agentic products with the vision to deploy AI agents in more than 2,000 branches. They will begin with complaints management, using Maestro as the orchestration layer and Autopilot to summarize and draft responses. From there, they plan to extend into other departments to deliver cost savings, higher customer satisfaction and greater efficiency.

Our Agentic product are not only deepening engagement within our installed base. They are also helping us win major new customers. A standout example is a 7-figure deal with a Fortune 15 global technology company. In a competitive win, they chose UiPath to power their SAP transformation and reinvent employee operations across functions like supply chain and manufacturing. They chose UiPath for our broad platform, including application testing, process intelligence and Agentic Automation and our proven ROI in complex enterprise environments.

What I consistently hear in conversations with our customers is the need for transparency and control. That's why we built Maestro to unify AI, automation, and human decision-making so outputs can be trusted with RPA playing a critical role alongside Agentic Automation. A great example is a leading U.S. health care provider that is piloting our Agentic products to streamline their accounts payable process. Agents monitor the mailbox, extract metadata, retrieve status updates and draft replies, while robots process attachments, send mails and create tickets.

At the same time, UiPath Maestro orchestrates the workflow, involving humans only for exceptions and final reviews, reducing manual effort by 60% and cutting processing time by up to 75%. Another strong example is a Fortune 500 American manufacturer, consumer and professional products who expanded to our Agentic products with a successful proof of concept in procurement operations. What was once a manual purchase order review is now handled by agents before a robot updates SAP. With UiPath Maestro orchestrating they cut latency and manual work while ensuring full policy compliance.

And we believe it's results like this that contributed to UiPath being recognized for the seventh consecutive year as a leader in the 2025 Gartner Magic Quadrant for Robotic Process Automation for both ability to execute and completeness of vision. As customers continue to scale their Agentic Automation programs, we are focused on helping them accelerate adoption. This quarter, we introduced 2 initiatives to support the journey. A team of deployed engineers to codevelop solutions and guide complex strategic implementations and UiPath Playground, a frictionless, no sign-up environment where customers and prospects can test, prebuild, verticalize agentic solutions and see the differentiated value of our platform.

Customers choose UiPath for our unified end-to-end platform that brings together API UI and AI-powered automation. This quarter, we are extending that advantage with API Workflows, our new foray into simplifying deployment and using APIs in our platform, giving customers greater control over how automation securely interacts with their business systems and data. This capability is essential to scaling agents securely and giving them a deterministic foundation to interact with enterprise systems.

Our focus is to continue building an open agentic ecosystem. While low code has made agent building accessible, growing complexity is driving demand for pro-code flexibility. That's why we launched UiPath Coded Agents. Now developers can fully customize and deploy secure and auditable agents built around their data and systems. A great example is Cato Networks, a cybersecurity company specializing in secure access service edge, who has deployed UiPath Coded Agents in production for IT ticket classification and analysis and is now piloting a master IT support agent expected to resolve up to 30% of tickets.

Following our initial integrations with LangChain and LangGraph, we have continued to expand support for other open source frameworks in the market, including LlamaIndex. This makes it even easier for developers to build in Llama and deploy directly into UiPath, leveraging our orchestration, security and monitoring for mission-critical processes. As enterprise automation evolves into agentic automation, the effectiveness of AI agents and automation depends on the quality of the data they can access.

To address this, we recently introduced UiPath Data Fabric, a unified data layer that powers agents, apps and workflows across the enterprise. By giving agents on automation access to trusted context-rich data, Data Fabric eliminates delays and ensure seamless integration across systems. Complementing this, our IXP capabilities help customers unlock unstructured information and turn complex documents into actionable data. This quarter, we launched IXP to general availability. And customers like Coronis Ajuba Solutions are already seeing an impact, automating data extraction from over 20 million pages and reducing errors by 35%.

Building on this success, Coronis plans to adopt Agent Builder and Maestro to scale denial processing across payers. UiPath's human-in-the-loop experience and flexibility to integrate custom LLMs are key drivers to support their enterprise-wide Agentic Automation vision.

The excitement around our agentic capabilities extends beyond customers with partners embracing them to create joint solutions that deliver greater value and faster transformation. A great proof point is one of our top GSIs. After leveraging our agentic products to automate order-to-cash collections, they highlighted UiPath Maestro as one of the most robust agentic orchestrators on the market. They are now committed to building over 20 agentic solutions across core finance process and claims and contract management and are already deploying them with joint customers like a global fintech company where initial POCs in corporate finance are focused on contract validation and revenue recognition.

Cognizant is another strong example of how our partners are embracing our Agentic products to co-create next-generation solutions. They are leveraging a variety of our solutions such as Maestro, Agent Builder, RPA, and IDP to build an intelligent claims processing suite. With this solution, we will jointly work with customers to reduce human involvement, streamline the intake process and de-silo processing across the entire claims value chain.

We also deepened our collaboration with Deloitte as they launched Agentic Global Business Services, a pioneering solution combining Deloitte's advancements in agentic AI with the UiPath platform to move enterprises from task automation to intelligent orchestration. Alongside our GSI partnerships, we continue to strengthen alliances with leading technology platforms. We are proud to expand our long-standing relationship with Microsoft who has reinforced UiPath as their preferred enterprise Agentic Automation platform for their customers. This strengthened alliance brings the power of Microsoft's industry-leading cloud and AI to customers through cutting-edge agentic products on the UiPath platform, like the UiPath Autopilot plug-in for Copilot and Teams and the bidirectional connector for Copilot Studio, which is growing in adoption and usage.

Lastly, in the U.S. public sector, with budgets now largely finalized, buying patterns are returning to a more normalized state and agencies are turning to automation and AI for mission-critical initiatives. The United States Navy is a great example with over 200 automations already deployed. The Navy expanded their IDP initiatives to streamline quarterly account reviews as part of its effort to achieve a clean audit opinion by 2028. As we look ahead, we have deep relationships across agencies, a strong team on the ground and we are well positioned.

Before I turn it over to Ashim, I'd like to invite you to UiPath FUSION, our flagship user conference taking place September 29 to October 2 in Las Vegas. FUSION brings together customers, partners and innovators from around the world, and we're excited to showcase how Agentic Automation is transforming enterprises and share more about our product vision and customer success. The future isn't about choosing between agents and automation. It's about combining them. Together, they are stronger and we are delivering best-in-class innovation in both categories to lead this next era of enterprise transformation for our customers and partners. Please reach out to our Investor Relations team for more information. With that, I'll turn the call over to Ashim.

Thank you, Daniel, and good afternoon, everyone. Before turning to the financials, I would like to reiterate the progress we've made on our operating priorities. First, on operational rigor and efficiency, we completed our restructuring and remain focused on driving productivity and disciplined execution. As a part of this, we've established a stronger cadence around predictability and key operating decisions, bringing product and field teams closer together to drive growth.

Second, on customer adoption, our teams, field organizations and partners are working hand-in-hand to deepen customer centricity and expand usage across our installed base. This includes strong momentum from partners leaning into our Agentic platform, such as Ashling Partners, calling it the natural evolution of RPA, and TQA underscoring how UiPath enabled their teams to help customers reimagine their operations and is driving record demand from both existing and new customers who want to lead their industries with game-changing automation strategies. These collaborations are critical to accelerating adoption at scale and moving customers from pilots to production deployments.

Lastly, while we continue to drive efficiencies across the business, we remain focused on investing in innovation as we look to capture the significant and expanding opportunity that automation and AI brings to the enterprise.

Turning to the quarter. Unless otherwise indicated, I will be discussing results on a non-GAAP basis and all growth rates are year-over-year. I also want to note that since we price and sell in local currency, fluctuations in FX rates impact results. Second quarter revenue grew to $362 million, an increase of 14%. Normalizing for the year-over-year FX tailwind of approximately $9 million, revenue grew 12%. ARR totaled $1.723 billion, an increase of 11%, driven by net new ARR of $31 million. Normalizing for the year-over-year FX tailwind of approximately $5 million, ARR grew 11%.

With the launch of our Agentic Automation platform, we continue to see customers moving to the cloud. We ended the quarter with more than $1.08 billion in cloud ARR, which includes both hybrid and SaaS, an increase of more than 25%. A great example is KLM Royal Dutch Airlines. After saving over 200,000 hours in 2024 with UiPath automation, they are migrating to the cloud, exploring Agentic Automation initiatives and implementing UiPath Test Cloud for SAP.

We ended the quarter with approximately 10,820 customers. As with prior quarters, the vast majority of customer attrition continues to be on the lower end. We continue to be successful in signing new enterprise logos that align with our strategy of targeting long-term customers with a propensity to invest, including new logos like Henry Schein, a Fortune 500 global health care solutions provider, who selected UiPath due to the breadth of our Agentic Automation platform capabilities and other notable logos in key sectors like the Watches of Switzerland Group, Community Financial Credit Union and the Vita Coco Company.

In addition to key customer adds in the quarter, our agentic capabilities and go-to-market improvements have resulted in deepening relationships in strategic cohorts. Customers with $100,000 or more in ARR increased to 2,432, while customers with $1 million or more in ARR increased to 320. Dollar-based gross retention remained best-in-class at 98%, and our dollar-based net retention rate remained at 108%, underscoring the durability of our customer base as they embrace our Agentic Automation solutions. Adjusting for foreign exchange, dollar-based net retention rate was 108%.

Remaining performance obligations increased to $1.209 billion, up 12%. Normalizing for the FX tailwind, which was an approximately $19 million tailwind, RPO grew 10%. Current RPO increased to $789 million, up 15%.

Turning to expenses. We delivered second quarter overall gross margins of 84% and software gross margin was 90%. Second quarter operating expenses were $243 million, a reduction of 6% from the prior year. Second quarter GAAP operating loss improved $83 million versus the prior year to $20 million and included $78 million of stock-based compensation expense. Our continued growth and disciplined expense management for cloud, operating expenses and stock-based compensation positions us well to achieve GAAP profitability in the near term.

Second quarter non-GAAP operating income was $62 million, representing a 17% margin, up more than 1,500 basis points year-over-year and driven by our continued focus on operational efficiencies. Second quarter non-GAAP adjusted free cash flow was $45 million. We ended the quarter with a healthy balance sheet of $1.5 billion in cash, cash equivalents and marketable securities and no debt. Our disciplined buyback activity reflects both confidence in our long-term opportunity and our ongoing commitment to return capital to shareholders. During the second quarter, we repurchased 8.3 million shares of our Class A common stock at an average price of $12.10.

Now turning to guidance. We are pleased with the team's execution in a variable macroeconomic environment, which is consistent with what we experienced over the last several quarters. And as Daniel mentioned, we are pleased with the progress of our public sector team. With this, we continue to maintain a prudent outlook and guide to what we see in front of us. Lastly, as a reminder, while we are encouraged by the early traction with our newly launched Agentic capabilities, adoption is still in its early phases, and as such, we don't expect a material top line contribution in fiscal 2026. Lastly, as a reminder, FX impact is recognized at the time of renewal and contract signing and is driven by European currencies and the yen.

Turning to the specifics of our guide. We are raising guidance for the progress we've made on our operating priorities and the incremental FX tailwind since we provided guidance on our first quarter earnings call. For the third fiscal quarter 2026, we expect revenue in the range of $390 million to $395 million. This range reflects an approximately $2 million tailwind, driven by FX rate movements since we provided guidance on our first quarter earnings call. ARR in the range of $1.771 billion to $1.776 billion. This range reflects an approximately $2 million tailwind driven by FX rate movements since we provided guidance on our first quarter earnings call.

Non-GAAP operating income of approximately $70 million. And we expect third quarter basic share count to be approximately 532 million shares. For the full fiscal year 2026, we expect revenue in the range of $1.571 billion to $1.576 billion. This range reflects an approximately $7 million tailwind driven by FX rate movements since we provided guidance on our first quarter earnings call. ARR in the range of $1.834 billion to $1.839 billion. This range reflects an approximately $7 million tailwind driven by FX rate movements since we provided guidance on our first quarter earnings call. Non-GAAP operating income of approximately $340 million.

And finally, we continue to expect fiscal year 2026 non-GAAP adjusted free cash flow of approximately $370 million and non-GAAP gross margin to be approximately 85%. Thank you for joining us today and we look forward to speaking with many of you during the quarter. With that, I will now turn the call over to the operator. Operator, please poll for questions.

[Operator Instructions] And our first question comes from Bryan Bergin with TD Cowen.

Wanted to ask as it relates to the client demand progression on your agentic solutions. Can you just talk about how the pacing is progressing in those accounts where they have moved from kind of POC and pilots to production? I don't know if I missed it but the mix of the clients that are developing agents here. And as you look at these accounts, I think you mentioned you're seeing agentic solutions increase deal sizes at a faster pace than the traditional RPA work in the past. Any context you could provide around that?

Yes. Bryan, let me start with the product overall, the Agentic product, and I will let Ashim comment more on the increased deal sizes. We launched our Agentic Orchestration and Agent Builder in May this year. And the progress that we are seeing is very encouraging. Having like 450 customers actively working with our technology, building agents with the intention to deploy in production is really a meaningful -- it has a meaningful impact.

I would say another good momentum for us is that most of those deals actually uncover even more opportunities for automation, scoring really on our strength, which is combining the orchestration, RPA plus API and the agent. And so I would say that the results so far are very encouraging. We've seen some significant deals that were driven by -- again, by the combination between agentic and automation.

Yes. And then, Bryan, I think when you look at what Daniel talked about, because people see the value in the Agentic platform, that is actually new product that we are monetizing, so it naturally increases the deal sizes for there. But what's interesting is the fact that it reinforces often the need for deterministic automation with our RPA and AI capabilities. So it has a twofold effect as they're going through their process transformations in our POCs as well as our pilots. That's really what's affecting our deal sizes.

Okay. All right, understood. Just a follow-up on DBNR, stable here 1Q to 2Q. Do you expect that to sustain as you go through the second half?

When you look at DBNR implied in our guidance, obviously, you can calculate and see what the normal ratio is. I would say we're stabilizing. That's what we feel. I won't make a commentary at this point to out quarters too far. But we feel, as we talked about in our guidance, we're assuming -- we continue to have a prudent outlook on the macroeconomic environment. We do see the government returning back to kind of normal buying behavior, which in my mind is positive. But with the macroeconomic condition, et cetera, we're still embedding the guidance as we did in the script.

Your next question comes from Jake Roberge with William Blair.

Great to hear about the nice start with some of your new Agentic solutions. There's obviously a lot of people talking about Agent Orchestration. When you're talking with customers about Maestro, what's the key pitch that gets you in the door? And who are you seeing most in those types of deals?

So our key pitch is being really agnostic. I think most of the customers right now are concerted into being completely on 1 side of a business platform. Because if you look at agentic and orchestration, most of the processes actually spend multiple business systems. So you will have to make a choice if you choose an orchestration or an agentic solution that is provided that one of the business systems.

How you are going to move data between business system? Where you are going to ultimately store your data? And many of our customers are really reluctant to choose for orchestration, 1 major business system, and they prefer an agnostic approach. Also, our Maestro is very tightly integrated with our automation platform. So it makes it extremely easy to combine agents and people-in-the-loop and actions that are provided by our robots.

And I think our platform is really the breadth of our platform in the terms of the offering of Orchestration, Automation and Agentic is one of the best in the market today, so this together makes it a very compelling offering for our customers. And we are seeing really interesting wins against major orchestration platform providers.

Okay, that's helpful. And then now that we're a year or so into some of the go-to-market changes that you all have made, how would you assess kind of the overall health of the go-to-market motion? I know things are always evolving, but do you feel like things are largely stable at this point in that motion?

Yes, I would say stable is a good word. I think we made really solid progress on making our entire go-to-market much closer to the customer. As we said, like in the past year, one of our major strategy was to be much more customer-centric and to break the silos. And I think our go-to-market, it's also working much closer with the product right now. And this is important in an era of where the iterations are largely driven by customer interactions. So we are pleased of how our go-to-market is structured and is functioning right now.

And your next question comes from Michael Turrin with Wells Fargo.

This is Austin Williams on for Michael. I just wanted to double-click on the U.S. federal business. And just any other color that you can add on how that business performed and just how you're navigating the uncertainty there.

We don't break out our segments, as you know. But I would say our public sector had just a really good quarter in terms of the momentum of selling with Agentic. We are encouraged by the progress in the public sector. The budget finalization, we see signs of stabilization there. The teams are really executing well so we feel we're well positioned for the second half. We just had a recent win with the Veterans Affairs and the Coast Guard. So we're really happy with both the wins in the quarter and the feedback we're getting from customers, including large customers like the Navy and the IRS as well.

Got it, helpful. And then just 1 follow-up on the big sequential step-up in subscription revenue this quarter. Was there anything specific to call out like related to that line? Is there anything onetime in nature that impacted that?

Yes. Going back to last quarter, we had kind of a leap year impact that was there that just kind of caught up in that, and so that was the kind of the movement that you saw. And now this quarter, it's back to stable.

And your next question comes from Raimo Lenschow with Barclays.

First, a quick number question for Ashim and then 1 for Daniel. The -- on the number side, if I look at your ARR guidance for the full year and FX and total revenue, given FX, so thank you for that, that's really helpful. You raised by a touch more than you beat. Can you speak to what's driving the confidence?

Yes. I think we're continuing to maintain a consistent philosophy. So when you look at the field sentiment, we look at what's in front of us in terms of the pipeline and just the momentum we are seeing with both the improvements that Daniel talked about on the go-to-market side as well as Agentic, Raimo, and that we wanted to reflect and really put that into the numbers for everybody to see.

Yes. Okay, perfect. That's really encouraging. And then, Daniel, one for you is like if you think about talking with customers at the moment, is there still kind of the market is still twofold -- like in 2 camps that you have some projects where people really understand RPA and process automation, et cetera, want to do that? Or is it all now like modeled together with agents, AI, et cetera? So in other words, is there like a core part of the business that can still continue to be sized? And then others is just kind of newer and hence, you have like you need to think about that? Or how is the market kind of behaving at the moment when you talk with customers?

Raimo, I think you can see the entire spectrum among our customers. There are -- in all fairness, there are customers that believe that AI agents will do everything. So they think very far-fetched in terms of swarm of agents that talk together. In the same time, I think I would say that majority of our customers are starting to realize that their automation programs are actually quite important to power their agentic initiatives.

I think it's becoming more clear in the market that the combination of orchestration, automation and agentic is really essential into delivering basically AI into predictable manner into enterprises. And as I said before, I think all the agentic exercises that we are seeing happening with our customers uncover more and more automation opportunities. They come -- typically, I can say they can come up with like 100 ideas that they call it agentic. And then when we look deeply, we discussed that 50 of them are better suited for automation.

Your next question comes from Matthew Hedberg with RBC Capital Markets.

This is Mike Richards on for Matt. I was kind of just wondering, now that we're more than a quarter in here, what the reception has been to the pricing of the Agentic portfolio? What have customers been saying? What are some of the learnings that you guys have had? Have you evolved that pricing as more customers have adopted the solution?

Yes. I think we are monetizing Agentic through a consumption-based model, which basically align very well with customer interest. I think one of the issues that is not UiPath-specific is the predictability of the pricing. I think everyone is trying to understand better how can you make a business case. We are working with quite a few of our customers to understand better this aspect of the business. But overall, I think the reaction to our Agentic pricing is positive and well understood.

Got it. And then just thinking through the go-to-market motion, you talked about bringing on specialized sellers. Is that specific for the Agentic solutions? Any incentives for the go-to-market motion to go after the Agentic opportunity? Or are you guys still sort of evolving that motion?

I would not say that we bring specialty sellers for our Agentic motion. We -- Agentic, our idea about agentic go-to-market is twofold. One is the horizontal agentic as part of our automation platform, basically completing our automation platform. So everyone in our go-to-market is equipped to deliver on this horizontal Agentic Orchestration and Agent Builder and the combination between Orchestration, Agentic and humans-in-the-loop.

At the same time, we have initiatives around vertical agents. Like we mentioned in the past, Peak is a great example where we have specialty sellers that deliver -- that are tasked with a dedicated task to basically inform customers about what that type of vertical solutions. And this motion is going to continue.

And your next question comes from Sanjit Singh with Morgan Stanley.

Ashim, I wanted to start with you. When I look at the guidance and I take a peek out to what that implies for Q4, if I sort of take the high end, it kind of implies a return to positive net new ARR growth, which I think for a lot of investors is kind of what we've been waiting for in terms of at least in terms of inflection from UiPath. I know you don't want to look out beyond the next quarter. But if you can maybe talk to the stability of the go-to-market organization and maybe the leadership that you've put in place. Are you confident enough that you've seen enough execution on the ground for multiple quarters in a row where that starts to become more likely in terms of seeing net start to inflect some flip from negative to positive?

Yes. I think -- let me answer it just also by looking in the context of third quarter. One is, I do want to acknowledge, like foreign exchange had some of that lift in there, which we acknowledged in our guidance, Sanjit, so I want to be -- make sure that we are transparent on that. But even when you look at third quarter versus second quarter, we're narrowing the year-over-year gap both operationally and then, obviously, some of the macroeconomic factors like foreign exchange helps that for us. So that's kind of 1 piece.

So you see the progression through the year, which frankly is kind of what we talked about at the beginning of the year, right, in terms of the federal government stabilizing, being really disrupted in the first half, stabilizing here in 3Q and 4Q as well as just the changes in momentum that we've seen in the go-to-market organization.

Specifically, am I seeing those signals within the go-to-market organization? The answer is yes. I think the commercial activity that we're seeing in pilots and POCs, it does 2 things. It reinforces -- it gives us opportunities to upsell, whether that's now or in the future, but it reinforces the importance of our platform and the architecture and the transformation journeys of a lot of our customers. And as Daniel said, we're really pleased with the execution that we're seeing in the ground.

Field, our field is empowered. They are customer-centric. We've reduced that bureaucracy. And we've really also closed the feedback loop between the field, product and management where we're able to react and really integrate with customers better. So all of that, combined with a consistent strategy of evaluating our pipeline, our data, et cetera, I feel very comfortable in terms of the guidance that we provided.

That's great color, Ashim. Daniel, for you, I had the opportunity to talk to some of the third-party industry analysts who we all sort of know, and they tell me 2 things. The first thing is that the real estate that UiPath is trying to occupy is where there's most value to be had and where there's value to be created. And the second thing they tell me that there's a lot of players trying to occupy some of that positioning where it comes not just being Agent Builder, but also being sort of the management layer, the orchestration layer.

And then the third thing they tell me is that this is really difficult to pull off. And so when I think about some of these 450 customers that are taking the leap with you guys on the Agentic, are there like initial processes that you are targeting, whether it's quote to cash or something around inventory or supply chain in terms of just trying to build that flywheel, build that trust, get those early successes? Do you have the sales and go-to-market organization sort of prioritizing specific processes that cut through various systems of records?

Yes, that's a great question. I think that we are both strategic and opportunistic in our approach. Indeed, we prioritize a few processes in health care, revenue cycle management, in particular, financial services, I would say, procure to pay, order to cash are areas of much interest to us. In claims management as well, we are seeing quite good movement.

But we are interested at this point to learn a lot about the blueprints of large-scale agentic deployment. I agree with you that the market is very complex at this point. And it's a great value to be extracted from this market. I think that we have some unique advantage because we are incumbent in this type of markets for a long time. As a reminder, we are in the business of automating manual processes since our inception. And most of the agentic initiatives are actually completing what we have started.

You just go beyond what RPA and automation was capable and just automate the steps of the process that couldn't automate before. But as an incumbent, having already robots that work as an action, it's a natural extension for us to add orchestration to provide better end-to-end process automation and agents that interact with our robots in order to get access to enterprise systems.

I think it's a really palpable advantage that we are seeing. So this is why with the product only a few months into production, we are seeing a lot of interest from our customers.

[Operator Instructions] Your next question comes from Brad Sills with Bank of America.

I wanted to ask about the partnership that you announced here with Deloitte. Is this indicative of just a greater focus on the SI channel as you embark on this agentic journey with company? Or is this more kind of indicative of a 3-way partnership with SAP and Deloitte? So just curious, really 2 parts to that question. One, what's the state of the SI channel? Is this indicative of a greater focus there as you get into agentic? And then the second part would be, what does this mean for your partnership with SAP?

I'd say this is indicative of both. Our 3-part partnership between us, SAP and Deloitte is bearing fruit and progressing really well. And at the same time, we are traditionally focused on extending our relationship with GSI. I think the major GSIs that we are talking to are basically in the process of understanding the market, making the bets on their platforms of choice for orchestration and agentic. And I'm happy to tell you that we are in many of these discussions, and we are seen as one of the major platforms they want to bet on.

Your next question comes from Scott Berg with Needham & Company.

This is Ian Black on for Scott Berg. Great quarter. Does the new Agentic portfolio enable you to implement RPA in additional workflows? Or is the focus on adding agentic capabilities to existing workflows right now?

We can extend our RPA and broader automation capabilities with Agentic. And as I said before, it's a renewed interest on our customers to do the exercise on identifying what processes are better suited for the overall initiative of bringing agentic and automation together. And we are seeing these initiatives surfacing more opportunities that we were seeing in the past years.

And your next question comes from Keith Bachman with BMO Capital Markets.

Ashim, I'll direct this to you. It looks like if I look at your ARR and the DBNR, it looks like you're still roughly 70% of your year-over-year growth is driven by existing customers and, call it, 30% for round numbers, from new logos. But your existing customers' growth is still down year-over-year pretty meaningfully. What causes that to flatten out as we look out over the next couple of quarters that your growth rate with existing customers improves? And the related -- my second question, a related question. You've clearly said on the call tonight that the agentic capabilities won't contribute to ARR this year but can it contribute next year in FY '27?

Yes. So let me take them 1 by 1. I'll actually just do the second 1 first. I want to be clear, like we said meaningfully contribute this year. We do see ARR agentic monetization happening and we are really pleased by that. And like I said when I answered a previous question, I think it also solidifies our position even in the renewal process in terms of the value of our form.

The second piece is, obviously, we believe with the momentum we have here, we see agentic to continue to contribute more and more. But I won't give specific guidance in terms of when and where and next year at this time. The question you asked about customer growth, for me, I'm assuming you're talking about the customer count. Remember that...

No. Sorry, the ARR from existing customers versus new logos, so really focusing on ARR from existing customers has been down pretty meaningfully over the last 3, 4 quarters. And I'm just wondering what caused that or when does that turn around.

Yes. I think we talked about this in terms of both the combination of the macroeconomic environment and government for the first half. And as you look at kind of our third and fourth quarter here, you can see that starting to come back from our vantage point, which is embedded within our guidance. And that's what we commented back all the way in March and it's played out very similarly.

Our existing customer base, again, if you look at the core metrics of customers greater than $100,000 and customers greater than $1 million, you actually see continued momentum in those cohorts of customers. And for us, that is very encouraging just as we navigate the macro environment and the stabilization of the U.S. public sector.

Your next question comes from Terry Tillman with Truist Securities.

This is Dominique Manansala on for Terry. So a variable macro environment just mentioned again this quarter. Just curious as to what patterns across geographies or verticals really persisted here in the second quarter. And are there any particular industries or regions where you're seeing more tailwind or areas of relative strength that you could be into? And maybe if you could just double-click on how your assumptions about the macro are really informing your prudent outlook for the rest of the year.

Yes. So when you look at pockets of strength, actually, the financial sector for us in the U.S. in health care, we really see that as pockets of strength right now in terms of customer demand and buying. The public sector, we're starting to see that momentum, as I talked about, both in the script and as in the commentary here. And the energy we're seeing from the public sector has been really encouraging.

That's been consistent across geos. And within Europe, I think some of our manufacturing customers also exhibit strength. When we talk about variable, it really is variable. I think it looks month-to-month, you kind of hear different things as companies are responding to different areas, whether that be tariffs or interest rates or geopolitical items. So variable really just -- it moves across the quarter and the year in a variable way. And I think that is felt not just by us but just if you open the news, you kind of see that in total.

And your next question comes from Devin Au with KeyBanc Capital Markets.

This is Devin on for Jason Celino today. Just 1 quick clarification question on the guidance, Ashim. Encouraging to hear that U.S. fed business has normalized. But does the ARR guidance raise still bake in prudence in that business or are you expecting incremental contributions from that segment in the second half, just given the more stabilized operating environment?

Sorry, can you repeat which segment when you said incremental?

Yes. Yes, the U.S. public sector, are you expecting, I guess, incremental contributions from that in the second half just because of the more stabilized environment?

Yes. So we see that normalizing, as Daniel mentioned, in terms of a more predictable buying behavior. I want to emphasize, I guide to what's in front of us. So as we comment, we do continue to bake in prudence. But at the same time, we're -- what's in front of us is a good amount of energy, and we obviously evaluate that with the tangible pipeline we're seeing from the U.S. -- the federal government at this time. So the answer is yes to both sides. We are baking in prudence, but at the same time, we do see more contribution from the U.S. public sector.

Thank you. And there are no further questions at this time. I'll hand it back to management for closing remarks.

Thank you, everybody, for the questions. And as usual, we are looking forward to meeting as many of you during the quarter. Thank you so much.

This concludes today's call. All parties may disconnect. Have a good day.

I would also like to point you to our safe harbor statement and remind you that today's discussion may contain forward-looking statements. Actual results may differ materially from these statements as a result of various risk factors, including those found in our SEC filings. We may disclose information related to development and plans for future products, features or enhancements, which are subject to change at our discretion without notice.

All statements are made only as of today, and UiPath undertakes no obligation to update any forward-looking statements and makes no assurances and obligation to update any forward -- assumes no responsibility to introduce future products, features or enhancements described today. Additionally, we would like to note that this is a product webinar, and as such, we will not be taking any financial-related questions.

In terms of today's agenda, we'll begin with a fireside chat with Daniel Dines, our Founder and Chief Executive Officer; which will be moderated by Hitesh Ramani, UiPath's Chief Accounting Officer and Deputy CFO. Following this, Graham Sheldon, UiPath's Chief Product Officer, will provide a product demonstration. After the product demonstration, if time permits, we will take questions from the audience. Please use the Q&A feature to submit your questions or if you're on a mobile device, please submit to the chat function.

And with that, I would like to turn it over to Hitesh.

Thanks, Jake. Daniel, how are you?

I'm great, Hitesh. And hello, everyone. Thank you so much for joining us.

Daniel, where are you dialing from today, by the way?

Well, I am dialing exactly from the places where we have started our journey, where we have built our first robots in Bucharest, Romania.

That's great. That's great. So let's get started. Daniel, we pivoted from an ACT 1 RPA company to ACT 2 Agentic Automation company. A lot has happened. Maybe it will be helpful if you can walk us through our journey towards Agentic Automation and why we believe it was such a natural next step for UiPath?

Since the beginning of our company, our vision was to play into end-to-end process automation. RPA was, in all fairness, just a really good gate into enterprises. It's very difficult to go and say, "I'm going to play directly into like API automation, system integrations. But RPA provided us an incredible easy gate to build great relationships with customers and be taken seriously.

We -- 10 years ago in this city, we were living 10 people in an apartment. So it was an incredible chance. But we were not limited to RPA throughout existence. We started with RPA, but then we extended into APIs. Because API automation is the -- it's one of the cornerstone of process automation. It cannot exist. One of our strengths is that we combine RPA, which means like using the user interface of applications like humans do with a very powerful world class API automation story.

And moreover, we moved into intelligent document processing because many, many processes start with the document, semi-structure documents. And then, we later moved into e-mails, which is form mostly of short messages. So we've built AI to understand short messages. And now, as you know, we've moved into understanding long complex documents. So Agentic is a natural extension, because -- and we call it second act, because all the other previous technologies, in fact, were rule-based, even if they were AI based. But the process could have been described in rules. And that was inherently limited, limiting the entire capability of our platform.

Now when LLMs arrived on the world stage, we seize an incredible opportunity to go after cognitive like processes. And I want to make a really important point. Our technology was based always to imitate how people work. What AI is doing when you apply in the context of processes, aims also to imitate how people work.

The way AI is going to make a decision, even if it's, I don't know, a loan application decision. The AI, it's going to emulate human's mind. So I think it's a very -- but the -- another big difference between how LLMs work and how our traditional automation work is that, automation is reliable and deterministic. AI is not-deterministic. So that was the biggest challenge that we actually had.

How can we take the power of GenAI and deploy it into a deterministic fashion? Here, we enter Agentic Automation. And we've built all of the frameworks in order to take the non-deterministic technology and apply it into deterministic processes.

So again, this is a natural extension for us that will enhance our robotic capabilities.

That's very impressive. In fact, I see when we say that we wanted to always emulate human behavior, I recall we were defining it as a digital FTE and the natural extension to that is now agents. That's very impressive. So now that we know why it is the natural extension to what we have been doing, what in your view is our right to win in the market, Daniel. And also, what are some of the use cases? How are we helping our customers solve some of their complex problems through this technology?

First of all, we are already there in the context of enterprise processes. I think many, many companies would love to have our expertise in understanding business processes -- manual business processes. Because in the end, if you -- where is going to be the most applicability of Agentic AI? Take an existing manual business process, emulate people and get all the benefits of AI, right? We are already in the context of -- we have 10,000 customers. Our robots work in the context of the business processes, we have the understanding.

Now for us, it's a matter of extending the robotic capability. So it's as simple as going to where are the processes, where you have robots. Let's look left and right and see what kind of task couldn't be automated before, let's bring agents that can help people. Actually, they will reduce the human input on these tasks and make these agents work with the existing robots. It's a natural extension. And if you think -- if you have already robots installed and you build agents, it's very handy to manage agents and robots within the same platform.

Because robotic, it's also an emulation of humans technology, we had to build a lot of security and governance that are specific for this type of technology. Nobody really has it at our level. This level of governance that is specific to emulation software.

We've been from the beginning, the way a robot access an application, how they handle people's password. You need to give robots passwords to access different applications. This is unusual for like workflow companies. They don't have to deal with this type of scenarios. But we have to.

So we are applying the same set of governance and security rules to agents. It's much more easier to handle and manage everything in one platform. It's like think about in your HR -- you don't have two HR systems, one for like, let's say, contractors, one for your full-time employees. You manage them on the same platform. It's normal. From the -- giving rights to your documents in the company, from everything, it's one platform. This is kind of the same with the robot.

So it's a natural right to be really there. And then, agents without actions are nothing. Most of the agents that we are seeing today are actually conversational agents. So it's chat-in, chat-out type of interfaces. So it's -- they don't run autonomously in the context of the business process. They are originated by human users that work on the specific tasks.

The agents that we want to deploy, and we are deploying are called and are instantiated by enterprise workflows. And they run autonomously in data centers in the cloud. And you have to put a lot more rules and governance around that technology in order to make sure that this is a reliable technology.

And again, we have the means to do it. And like people, when an agent naturally touches multiple systems. If you look into an enterprise context, it's very rare. When a business user lives only inside one business application, it's very rare you live only in Salesforce or in SAP. You connect to multiple business system in order to carry on even most of the tasks. That means that the actions that agents need in order to complete their goal, touches multiple systems. This is where we shine.

We have the Switzerland of integrations with different systems. We are agnostic. We are not going to provide better support for Salesforce or better support for SAP or better support for dynamics or Oracle or whatever. We will be agnostic and we will provide equally great integration support to all platforms.

So it's natural. I heard it from CIOs of big healthcare company, financial companies saying, yes, I'm not going to put my data from one system into the other in order to power the agents. I like your approach that is agnostic. And I like your capabilities to integrate all systems. It's -- in the end, why RPA exists? Because there are systems, where it's very difficult to communicate to it, because maybe they don't have APIs. But maybe they are very complex. API is very difficult to implement that require high-level expertise.

Where there are legacy systems, we have mainframes in banking system, financial system, we routinely automate tons of mainframe application. I don't think they will disappear. They stay for 30 years. I don't think they will disappear in the next year.

And we are really best-in-class when we connect to these systems. In many enterprises, agents will connect to multiple systems, some legacy systems, some modern API systems, but it makes sense to have a platform that handles everything.

Why would you handle? Why you take a platform that handles APIs? And then you have a different platform when you'll need RPA. It makes no sense. It creates a lot of security imbalances. It's much more difficult to manage the security. So it's a natural really way for us to expand.

And you asked me also about use cases. We're seeing really use cases across the board. And look, we've been always into the finance department, order to cash, procure to pay huge processes. Now customers have the chance to get an amazing level of automation. We are using with one of our largest customer in Japan, a large bank in Japan. They want to achieve in their order-to-cash process, 95% accuracy. It's a huge number, and they started with automation around under 50%. So it's a great opportunity.

We are seeing in healthcare in the context of revenue cycle management. We are working with many customers, client claims denials, prior authorizations. And then, it's -- in customer service, obviously, there are a lot of use cases and many more.

And I think, Hitesh, you are in charge with our own internal automation. Maybe you can elaborate a bit what we are doing internally?

Yes. No, absolutely. And Daniel, as you mentioned, we started with first transforming our finance processes using automation as part of our ACT 1 and there were several areas where we could not actually solve some of our pain points, which we are now actually using Agentic solutions to help us. And I'll give you an example. Within our order-to-cash process as a company itself, there were still manual reviews of contracts that the revenue recognition team was performing.

Now we are using our IXP technology along with Agentic. So we can actually transform how we are reviewing these contracts by putting human in the loop. So in my mind, there are two things that has actually helped us really lay a strong foundation for autonomous workflows. One is the fact that the solutions which we could not solve in past, now with Agentic, we are able to solve it. But using Agentic along with robots and Orchestrator, we are actually able to embark on our autonomous journey. So I feel like we are at a very great spot in terms of the evolution of technology.

Maybe with that, Daniel, it will be helpful, I guess, for the folks on the call, if you can explain how important Orchestrator is for Agentic workflows?

From our perspective, this is the essential component in order to deploy enterprise agents. And let me explain why. And we thought a lot about it. Our foray into Agentic AI started actually a year ago.

I we thought hard, how can we deploy these agents? And we realized -- first of all, our blueprint is agents will not be allowed by our enterprise customers to take directly actions that can provide important side effects and cost security issues were, because I know people will not let agents move money by -- just by agents or maybe in very small quantities.

So we realized, we need the framework to have humans in the loop, a very powerful framework. And we need a framework that will allow agents to connect with agents and agents to connect with robots. And we did a great, or we call it an orchestration engine. RPA had an orchestration engine, but it was not adopted to the model to the way that -- the modern way to describe processes and to facilitate this integration.

So we created from scratch a new orchestrate engine. And I think we're really were ahead of the market. A year ago, nobody was even talking about Agentic AI. And nobody was talking about orchestration. Now everybody is talking about orchestration. Because everybody realized that you need orchestration in order to deliver agents.

And every more -- this is -- orchestration emulates also how human organizes in groups. If you think it's -- there is a reason why we create processes, why we create workflows that are governed by rules and by fixed path. And it's mostly the individual tasks that one person is doing that required knowledge and cognitive approach. But the end-to-end workflow, it's more than a fixed path. So you need to have this engine. And you need to have an engine that is modern and is capable of connecting agent to agents or agents to robots and putting humans in the loop.

So we invested a lot in this technology. And we -- it's not -- it's an orchestration means, orchestrating tons of workflows. A process you can think of business process, end-to-end business process is comprising hundreds of sub workflows. We have great analytics that makes like a 360 view of all your process instances that comprise agents and robots and automations, APIs, document. Everything is fully auditable. You see one single audit trail for everything.

Even our process mining technology can connect to back-end systems and we can integrate these data sources. So you can completely have a 360 view of the transactions in the system. So again, this resonates with our customers.

I'm explaining to customers, yes, this is how we are seeing -- what do you think? Do you think we are wrong? Do you really believe -- are you ready to deploy a swarm of agents that you don't -- is not clear how they talk to each other and end of the -- and somehow magically they will deliver the results. It's not. People want this reliable way to deliver agents. And that's the -- this is what gives comfort to enterprise customers.

You put humans in the loop in orchestration, you watch how the agents perform, people approve. Agents become actually better and better every time they see, they learn from how people interact with them. And then once you become more comfortable, you put -- you surround them with more -- with rules, but give them more agency. Like I am comfortable that in case of a loan originating for loans less than, I don't know, x amount of money that came with people with this credit score and that different characteristics, I can bypass humans.

Yes, but only after I have -- I need to have the orchestration, first of all. I need to have the audit. I need to see. I need to have the confidence. This is why this is tremendously important. This is why we are winning deals today, because our orchestration is really thought for the Agentic era.

Yes. No, that's helpful, Daniel. Now that, Daniel, we have GA-ed our Agentic products almost 3 months. And I also am aware that we had several of our customers, which have taken part in a private preview and since then they've been actively working on use cases. What has our early feedback that you are hearing back from our customers, the early feedback on the products that they've started using. And also, where do you see them evolve like go from here on forward?

Even before we launched our product in GA, we had customers asking us to GA, because they wanted to put in production some of the agents and orchestrations.

For like a few of our customers, we had a controlled GA. So we gave them the assurance that we are going to support the product even before being GA. So they put it in production. It's the highest number of POC and pilots since our early days of RPA. And it's a higher number. We are a much bigger company, and this is our bottleneck of really fulfilling the POCs and pilots requests from the customer.

So the interest of this technology is tremendous. But it's still -- I want to make clear, it's still early days. Customers need to get confidence. Trust is a very big word here. They -- many customers still start small. They put agents into the small part of the process into the small context in order to get the confidence. We and our partners and customers are in the ways of figuring out what are the blueprints of large-scale deployments.

My estimation is that throughout this year, this trend will continue, but we will learn a lot more. We are building more and more vertical solutions. Agentic, it's very suitable to build more vertical solutions because in a way, I want to describe to you a bit how I see an agent. It's more like a college grad that goes to their first job. They know really the public space. They have a lot of knowledge, but they don't know their industry. They don't know any company specific.

So if I go into banking, I need to learn a lot about banking specific. If I'm becoming an investment banker, I need to learn about so -- but this is an industry specific. And if I go to a specific bank, I need to understand the specific city of that bank.

As we work with our customers, we capture a lot of industry knowledge. We already have great industry knowledge from our RPA. Now we are working on capture more industry knowledge. We come with better industry solutions that we can replicate to different customers. So we are in the process right now to understanding, again, what is the blueprint of large-scale deployments.

That's very helpful, Daniel, thank you so much for your insights. This was valuable. With that, let me turn over the call to Graham, our Chief Product Officer, who will provide us a product demonstration.

Great. Thank you. Good morning or good afternoon everyone. Thank you for joining us today. I'm Graham Sheldon, the Chief Product Officer at UiPath. And I'm really excited to walk you through how our customers are actually able to unlock AI transformation and drive real ROI time to value through our platform.

I'm going to start this morning by giving you a view of our platform and how we've reimagined it to -- in this new ACT 2 that Daniel described around Agentic automation.

There we go. The UiPath platform that we're building is built on our strong foundation in automation. Our customers can now easily combine the best of what robots have been able to do about deterministic rules-based kinds of work with what agents and LLMs are now capable of doing, the more dynamic and goal-seeking type of work. And it's the combination of that in an end-to-end process that can really help transform the way that we do business and the way that work gets done.

People are still at the heart of this, and they can really then focus on the critical decisions and the high-value work. And that together runs seamlessly across all of the workflows, all of the systems, all of the people in an organization in an enterprise-grade governance framework, so that you can feel confident and have the trust that those processes are running the way that you want to and you remain in control, fully integrated into your systems regardless of whether they be legacy ones or newer ones as we move forward together.

So why is this platform uniquely able to deliver these AI transformation results for our customers? First and foremost, we believe in what Daniel described for these mission-critical workflows, you really want to have specialized agents, not ones that are swarms of agents that are going to go run around and do things that you can't understand or control. And our customers demand that they have real insight into what those agents are doing, specialized for particular tasks so that you can trust that they're working in a way orchestrated across the workflows with what people need to do and what robots will need to do, so that you get the highest level of accuracy, reliability and governance.

Those agents are only as good though as the actions that they take. And those actions are based on the best-in-class automations and tools that they have. Again, over any system or any type of data, we've built into the platform those capabilities, and we believe that they remain best-in-class. That's work that's done that's orchestrated with people and with robots to be able to operate across an entire workflow through end-to-end agentic orchestration, which I'll show you in a few minutes.

The UiPath platform supports pro code, so professional developer tools, and those can come from any vendor, including our own, so that you never get locked in and you can use the best-in-class models, best-in-class agents from any provider. But we also have low-code approach so that it can scale out and you can see -- you can have your business users and domain experts participate in the development of those agents.

And underpinning all of that, we're investing a ton into making sure that you can track what those agents are doing, govern them and manage them, and they achieve the highest levels of accuracy and reliability. Now we've developed a ton of new innovation on the platform, but I want to focus on a few highlights that take these promises and make them a reality for our customers.

First is a product we call UiPath Maestro. So this is that concept that Daniel described in terms of agentic orchestration. With Maestro, we're not just simplifying the automation itself, but actually how teams manage it. It's actually a collaborative surface where you can design, you can manage and then you can optimize these workflows and plug-in agents where -- for the right tool for the right job, plug-in robots for the right places and allow people to participate as well, orchestrating all of that seamlessly in our system.

Next is Agent Builder. Agent Builder is a low-code interface for designing these specialized agents and has an integrated experience to help you test and evaluate that over time so that you get the highest level of accuracy and reliability from those agents.

And last but not least, we know there's a ton of data that's still locked away in unstructured or semi-structured data like documents and communications. And IXP is a product that now allows you to sort of extract and make that data really useful in these end-to-end automations.

And all of these capabilities come together. I'm going to show that to you in just a moment on the platform. And so, this platform is trusted by over 10,000 organizations worldwide. And we have customers like WEX that are using these agents to tailor customer profiles and improve sales engagement. You've seen lots and tons of examples of this across many different industries, many different departments to truly transform the way people work and get incredible results.

Now I'm going to focus in on one example of that in the context of insurance claims processing. So in this example, today, people are doing a lot of manual works still. You're gathering critical information, you're doing some research, you're updating systems of record and systems of engagement, communicating back and forth with the claimant. And it's a really difficult process. It's very messy. There's a lot of swivel chairing happening.

And think about that daily reality for an adjuster, with multiple legacy systems simultaneously open, and it's really hard to stay on top of this. It's also fairly error prone. And those mistakes cost time, but they also cost patient outcomes. And that's what we want to make sure that we can really transform.

Now imagine a transformed experience where instead of that adjuster doing all of this busy work, instead of that, we give -- agents and robots take care of all of that sort of repetitive paperwork. And they can focus really on just the part where the final decision needs to be made, where the agent has already made a recommendation on the basis of all of the data that it's extracted, all of the policies that's been applied and they can just do the last mile, the critical decision about whether or not to move forward and any additional information that can help this particular transaction go as quickly and as seamlessly as possible with natural language tools to help them get their job done even more efficiently.

So how do we make an experience like this one, a reality? What I'm going to do is now show you a demo, but I want to start with the process. The process in a very simplified way, again, involves a lot of manual work. And with RPA and automation today, robots are already doing a lot of this work, getting the critical information, making some simple rules-based identification of the claimant and then updating the systems of record.

Well, now with UiPath Agentic automation innovation, we can now take out some of the rest of this. You can take out a lot of these sort of more nuanced and dynamic decisions and make recommendations to people like doing the initial duplicate claim check or checking for fraud as well as a determination of eligibility. And then, you want to orchestrate this process end to end. This is obviously an oversimplified view of this process. The process actually is quite a lot more complicated than that.

And a lot of the time to value is really unlocked when you can describe this in a way that lets your developers and your business users collaborate together to make this a reality.

So now I'm going to switch and show you what that looks like in the UiPath platform. Imagine this particular process is now something you can model directly with UiPath Maestro. I'm going to switch over and show you what that looks like in our platform.

And here you go. This is that same HSA process working end-to-end in the UiPath platform. And as I mentioned, you have some robot tasks where you're gathering critical information, you're updating systems of record. You also now have agents doing eligibility checks. We even have LangChain agents here built by professional developers doing a fraud check as well as ones from other systems and other agent providers like this one from Agent force that might interact and notify the claimant, all orchestrated together.

And all of the actions that those agents are taking, all of the things that the humans are doing only when they are needed to are tracked, which is super important to make sure that this traceability is happening. So in this case, if we've got an eligibility determination, and we want to see exactly why the agent made the decision that it did, which in this case is to recommend that we approve the claim. I can go back and I can see all of the different information that it gathered, all of the different reasoning that it did tracked here in an auditable and traceable way.

Now many of these processes, as I mentioned, have a lot of documents that you're gathering from the claimant. And Maestro is going -- is helping me to weave these together. Let me show you how these documents are processed in our system. In this insurance scenario, think of all that unstructured data that I was here, and let me just provide my meeting controls for a second. There we go.

In IXP, all of those documents, the PDFs that contain the claim summary, the explanation of benefits and the each of the individual claim details are being captured by LLMs. And with simple prompts, we're able to get all of that information from structured documents, even unstructured ones like the receipt here. And I can actually go into IXP and with just -- and much, much quicker than you ever able to do before. You can see visually where it's extracting some of that data from and be able to check and see how well this model is performing over time.

Now that we've seen how effectively IXP can extract some of that critical data from complex documents and communications, let's go and check about the insights seamlessly feed into the broader automation workflow.

Going back to Maestro, I'm now going to focus on this agent and how we built it. This agent was built in Studio as a low-code agent for determining eligibility. And you can see that we've got some critical elements. We've got the prompt or the steps that the agent should take in order to figure out whether or not we should move forward. It's got the tools, existing automations, web search tools, and we're going to support things like -- we do support things like MCP for tools and agent-to-agent communication.

So you can reuse all of your existing automations or you can use even other agents as part of the definition for this agent as well as the critical context from your policies and how to escalate to humans if the agent gets stuck.

I mentioned before, we're really focusing on reliability and accuracy. And in line here, it's kind of hard sometimes for people to understand how well their agent is doing and the improvements that they should make. I didn't come up with this prompt myself. I got help from Autopilot actually. And I asked how to make it more concise. And I can continuously look at the suggestions that are here in this particular case, maybe to make this a little more concise. And this loop helps me create the best possible agent.

But how do I prove that? I need to be able to evaluate how this agent is performing both right now and then over time to make sure it continues to meet my quality expectations. So I can look at results from many different evaluation sets. And for every one of these, I can go deeper and say, okay, in these particular cases, why did the agent make the decisions that it did and be able to compare with these test cases what I expected versus what the agent actually did and then be able to update it and make recommendations back to the agent so that it can perform better next time.

It's this integrated experience that we think is really unique to creating specialized agents in this enterprise context that you can trust.

Now the last piece I want to show you is that we've just designed this great -- this process. And you can see that it's sort of paused at a step where I'm going to make a final determination about whether or not to proceed with this claim.

In the UiPath system, we have an integrated way to build an action app, and I can bring to the end user the final determination that they can make a decision on. So as a business end user, I can take a look at the analysis that the agent provide with the summary. I can inspect the documents myself to make sure that indeed, it's making the right decisions. And I can provide additional comments so that over time, the agent can get better and better with long-term memory.

Last but not least, we know that this process is an evolving process. So we want to control the degree of agency and over time improve it. Well, with the UiPath platform, you can also look at how this processes going over time and be able to show where there are bottlenecks, where there's conformance problems and even go so far as to try to figure out how to optimize it, to add additional agents to see where conformance is not happening and to be able to simulate and rework this to eliminate those bottlenecks to improve this and see where I have additional ways to improve this process.

So to summarize, Maestro is more than just modeling. It's a central command center where agents, where robots, where people are collaborating together in a very highly governed, highly accurate and trusted way so that for our customers, processes that now took weeks and months take days. You're eliminating all of that busy work. You're focusing people on the stuff that really matters.

And for UiPath, it's about this our core differentiator, bringing the best of what robots can do for deterministic work with the best of what agents can do and help your employees focus on the stuff that really matters to be more productive, more confident and to satisfy your customers and hopefully create better health outcomes in this particular case.

So that's the future of Agentic automation with just a snapshot of what is possible today with UiPath on the Agentic Automation platform.

Hitesh, I'm going to turn things back over to you.

Great. Thanks. Thanks, Graham. This is really cool and very exciting, Graham, to see how automation agents are operating along with humans on the site, and it's all getting agent orchestrated with UiPath Maestro.

Let's take next 10 minutes to address some of the questions. We are seeing a few questions come through the Q&A here. If folks on the call have more questions, feel free to ask them.

Let's get started with the first question here for Daniel. The question is, do you need to see broad enterprise adoption before seeing agency orchestration adoption? Why or why not? It sounds like I think folks want to understand, is there a sequencing? Does the enterprises need to do first embark on Agentic adoption and then focus on orchestration? Or can that be done side by side?

I think that for the context of autonomous agents that are deployed in the context of enterprise processes that large -- that run largely autonomously in the back end -- back office processes, you need simultaneously orchestration and agents. We believe there is no other way to deploy them. And this is the actually something that I said before resonates deeply with our customers.

Great. Now we saw in Graham's demonstration, Graham, you mentioned about there was a LangChain agent. And Daniel, you mentioned about how we are Switzerland. What do we mean by this? And how UiPath Maestro is expected to interact with other agents?

Maybe, Daniel, if you want to answer that question?

So we thought a lot, how -- where the world is going in terms of building an agent? And while we have tremendous experience in building low-code applications, low-code RPA. We know that also for sophisticated scenarios, people rely on code. So we decided to have a dual approach on building agents to have our own low-code agent builder that really can speed up the development, but also to support most common Agentic frameworks, open source frameworks like LangChain and LlamaIndex and CrewAI and there will be others.

Because in the end, we want to -- always we said we are an open platform, so we want to offer our customers the ability to build agents in whatever flavor they want. But even with this open source frameworks, we offer the same level of integration with Maestro, but also with our robots as actions. So you can create an agent in LangChain. You can easily describe robots as tools for the agent. And then you can back the agent, you can send it to our orchestrator. And then you can see it as an agent in Maestro exactly like you see our own agent, same security and governance applied to it.

Moreover, we -- when building an agent we realized that one of the most complicated area is testing the agent. And we also have tremendous experience as a testing platform. So we focused a lot on building an evaluation sets for agents and helping our customers improve the prompts of agents. This set of technologies apply to our agents, but to apply also to agents develop with the open source frameworks. So we treat them literally as first-class citizens in our platform.

Awesome. So I see there is another question. Given the current rate of change or evolution in technology, why wouldn't the agents eventually be both deterministic and probabilistic and completely replace any need for deterministic robots?

Yes, I think this is a very important question. And it's -- because you can say humans, yes, can do both cognitive and also rule-based tasks, I would say. Why agents cannot do this? I think it's related to the existing limitation of technology.

GenAI is extremely good finding patterns in data. So you can scan tons. It seems a lot of data. But GenAI doesn't follow rules in a traditional human sense. It's very difficult to give a set of steps and rules and have GenAI follow these rules precisely every time. It's not gone -- it's going to deviate from the rules.

So I've seen even recently, I've seen an interesting study published by researchers of Apple, where they try to convince GenAI to follow a simple algorithm for this problem of Towers of Hanoi, that is moving this from poles to poles in a certain order.

GenAI, even if they said I can give you exactly the step-by-step, the algorithm, GenAI cannot follow it. So that's the reality of the technology today. We also -- another point is, when you try to -- it's more difficult to build something in plain English than to build something that is reliable in a programming language, especially in the local language like what we have in our studio.

In English, it's much more difficult to reason about every single change that even a coma can change the outcome of the agent, changing from an LLM to the other can change the outcome. You never are capable of testing comprehensively testing an agent, because it's predict -- in code you can go through all the branches and you can have a fair understanding of how the code works.

It's easier to train a developer to follow to describe a rule-based process in code than to create an agent. So I -- and that's clear to me. Why would you use -- and even philosophically, why you use a technology that is a sophisticated to do something that is rule-based.

Yes. I mean it's like, for -- in my viewpoint, Daniel, why would you hire a PhD to perform a clerical task, let's say, that way, number one. And the other thing is when we talk about governance and trust layer for me, it's very important to always get a consistent answer. So if there are certain automations that will give me a consistent answer every time, I would rather like to use that versus an Agentic. And as you said, I think Agentic, it's not guaranteed that every time we get the similar consistent answer.

So I see there is one more question out there, which is -- how do -- when we talk to our customers, how do we talk to them about ROI with regard to orchestrator, UiPath Maestro? And how difficult it is for our customers to build their own orchestration platform like Maestro?

Well, building a platform like Maestro, it requires years of engineering and teams of hundreds of people, Hitesh. This is not something that is easily achievable. First of all, I want to make sure that we -- the audience understand, we've build Maestro on the top of our existing orchestrator. We start building our orchestrator around 2015. And I think it's achieved maturity probably around 2020.

On the top of our orchestrator that can provide the security and the governance and deployment and managing of robots, we've built the Maestro, which is more like a workflow engine, but it's combining. So building these capabilities are extremely difficult. And it's -- there is no point into doing this. It's -- in this way, you can start building from scratch any other. So it applies built and buy to like most of the software, but it's not a simple system. This is an extremely complex enterprise system.

Yes. And Daniel, as I think about ROI from my viewpoint, RPA by itself or agents by itself, or maybe orchestrated by itself has relatively limited value. But when you integrate all of them together, along with the ability to go and play the Switzerland role, in my mind, that actually lays the strongest foundation for autonomous workflows. And I feel like that has the maximum ROI from the way.

Yes, because, Hitesh, end of the day, what customers want is outcomes. And we provide the outcomes. You cannot have random agents deployed and spread around. You need to put them in the context of an end-to-end process with the workflow, with orchestration, with humans in the loop. And in the end, that provides the outcome.

I think the other thing maybe perhaps that's worthy mention is time to value. So yes, in theory, could someone build -- there are basic frameworks that help you string together agent work or some robotic work. But it does not have the governance, the observability that controls the evaluation sets, the ability for you to inspect and then collaborate on it together. Those are things that the scarcest resource in most of our customers today are the data scientists.

People just don't have a lot of folks ready to go build agents themselves. And so, to have a surface where you can have developers and business end users looking at that diagram together, being able to figure out what needs to be built and then getting it into production within days. We're talking about days for some of our biggest customers. One of the biggest healthcare providers was able to turn around that claims scenario just like the one we did in 48 hours.

Yes. That's great. I guess the other question is, we recently acquired Peak. And what are some of our learnings from this acquisition? And how do we expect our vertical solutions to evolve over time?

Peak is, -- it's looking like a really good acquisition for us. It's well received by customers. I think, we can accelerate -- we are seeing an acceleration of their pipeline. And our go-to-market is excited about adopting big, especially in our manufacturing verticals, mostly in U.S. but also in EMEA, Germany, in particular. And it's also a lot of lesson learned from Peak.

Peak use the model like forward deployed engineers in order to facilitate the Agentic. This is a model that we are looking more and more to adopt ourselves into helping our customers in the early phase of Agentic and we will use their model to also build other specific vertical agents. We already identified a few areas, and we are putting in place similar teams to build the technology.

Yes. Daniel, there's one more question here. For customers who are using other RPA technologies and dimensions such as Blue Prism for a minute. What do you think are the key advantages of our offering versus theirs?

I think, we are seeing a bit of an acceleration of customers that want to migrate their RPA solutions. But Blue Prism in particular, I think, has not evolved in the last quite a few years. So it's -- I always said it's a good solid platform for RPA alone. But when you try to extend it, it's a very difficult platform to work with.

So now customers have realized that the importance of having an overarching platform. This, again, I want to make a case, combining many agents will need RPA robots. Managing agents and robots in the same platform offer tremendous advantages. And you can have an existing center of excellence is working. It's an extension. It's a natural extension of your already automation programs in place. So this is why some customers are saying, well, maybe this is the end of the road with this particular technology. Why we are not switching to a technology that offer me a future.

Great. Well, I think we are at top of the hour here. Graham, thank you so much for the demonstration. And Daniel, thank you so much for your insights. It was very valuable. Do you have -- Daniel, do you have any closing remarks for the audience on the call?

I would like to thank everyone for their time and for being with us today. We are looking, as always, to connect with many of you throughout the quarter. We are already here maybe to provide more information and everything you guys need. Thank you so much.

Thank you.

All right. I'm Siti Panigrahi, software analyst here at Mizuho. It's a great pleasure to welcome Ashim Gupta, COO and CFO, 2 hats. Yes. That's great. Welcome you to the conference.

Thanks so much, Siti.

I'll start with a pretty -- you reported recently, fiscal Q1, it was much -- better than expected. So maybe why don't you start with your puts and takes for Q1, where you saw the strength, where you saw the weakness, let's start with that.

So look, I think Q1, we delivered revenue very strong. We beat consensus estimates, we beat consensus estimates on profitability. You can see strong free cash flow generation of $100 million or more. Net new ARR just above what we -- a good beat versus what we had also guided in March. And we did that in the backdrop of what I think we can all agree is very uncertain and kind of volatile environment that we've been facing.

So we're really pleased and we're super pleased by the execution with the team. I think beyond the numbers, when you go to in the financial statements, when you just go to what the team has executed, new partner program in place, which was really well received.

Strong contacts and agreements with GSIs, good -- like Deloitte for Agentic ERP, really good momentum, but it was underscored and highlighted by the launch of our Agentic platform. So I know in the world of AI, a lot of people, there's a question of like, what's real, what's hype, et cetera. We launched -- we had the most significant product launch for us, tangible software that our customers had been previewing the previous 3 months, that is now in GA, that allows them to build and deploy agents at scale in conjunction with the rest of our automation platform. And then also Agentic Orchestration, Agentic Testing and Agentic Orchestration, allowing humans, robots and agents to be working together and really bringing that to our customers.

Definitely, I want to spend more time on the AI side, Ashim. But one other question I was getting post your earnings is FX. How do you factor the FX impact and how that -- how you look into that in your guidance?

Yes. So I think if you just step back, the first thing is we've been very consistent about the way we report and the way we guide with foreign exchange. When you go back in history, we really haven't seen as much volatility in foreign exchange, so this hasn't come up. But when we reported in March when we guided for the year, the euro, which is the most exposure that we would have is -- was hovering between $1.08 and $1.09.

When we guided in the second quarter or at the end of the first quarter, it was around $1.12. So it was very minimal. If you go back to September of last year, when we guided in the third quarter, the euro was around $1.07, and we went into December, it was more like $1.02, $1.03. These movements always happen. And we never adjust our guidance for what we see as immaterial amounts.

What really confused people, I think, is when you look at other software companies, they talked about material movements. So if you had done your earnings a week before us in March, you were dealing with a euro that was $1.02 versus $1.08 or $1.09. So it was a much more material impact.

The last point is within our ARR balance, more than 55% of our ARR is in U.S. dollars. So then you take a smaller portion that is in euro. And then that -- when you carve that down with contracts that have multiyear exposure, which is a good chunk of our larger dollars, you have an immaterial impact of contracts that are renewing this year itself.

And the other -- the last point that we wanted to make is we don't mark-to-market our contracts every quarter. So we will only adjust the foreign currency and realize the currency loss or gain when the contract comes up for renewal. So those are kind of 3 points to give -- to put that in context.

Thanks for clarifying that. The other point is, at this point, macro. We are seeing so much variability there right now. So how does that play out for you? Are you seeing some kind of deal elongation? Or are you seeing the smaller deal size? Anything you are seeing?

Yes. I think it's been across the board. So it's -- I would say 2 ways to answer it. The first is in March itself, that is when we really recognized and called into our number, macro uncertainty that's there. I think when you turn on your TV or you open the newspaper for anybody who reads a newspaper still or you turn on your iPhone, uncertainty still is very prevalent within the macro environment. So we don't see that as fundamentally changing.

That does affect deal cycles. Why? I think customers are still trying to understand what is their revenue and cost structure going to look like in the environment in which we're in. And so I think we saw a little bit more clarity from certain customers come in the second quarter, come at the end of the first quarter, which allowed us to adjust our revenue up in the second quarter by $12 million plus.

So when you look at that together, I think the overall environment remains uncertain, but we're hopeful that, that clarity comes within the second half of the year. That said, in our total year guidance, we remain prudent, but this environment is going to continue for the near future.

Yes. Another topic was federal. I think you talked about some kind of cautiousness on the Q4 call. So then how do you characterize the federal business? What's kind of baked into your guidance?

So I think the federal business is going through transition. I was down -- we've talked to -- I personally have talked to a number of the undersecretaries as well as several of the CFOs of the different agencies. Many of them are just going through the transition. They don't have their budgets nailed down in many places yet. They're still kind of working through DOGE's control, influence, what DOGE wants to review, et cetera.

And then the third piece is there's just a general sense of uncertainty, even some of the areas where even the heads of an agency has either not been confirmed or is in the process of getting confirmation right now. So what we factored into our guidance is really a more renewal oriented deal flow for the first half. We do see the second half with budgets getting confirmed in September. There should be some life into the new business.

That said, when we talk to our customers, UiPath is very well positioned. Department of Government Efficiency, they're not hiring back the employees that were let go. So they're trying to contend with how do you do the same amount of work with less people. And that is where both robots, agents and overall automation thrives. And we're very aligned with them on the vision and the opportunity together.

Okay. Now with that, let's switch to strategy. I mean, going back to 2016, UiPath was a leader in RPA, you have pretty good success there. Now coming to Agentic AI. Now this is something like a second act for UiPath. So help us understand what's your strategy there? And what are you hearing from customer?

Yes. The first thing is, I think we haven't been an RPA company since 2016. I think I -- that message I really want to hit home. We extended our platform to include Document Understanding, Test Automation, process mining, communications mining, which is the mining of e-mails to identify and drive efficiency and create routing. We had done all of that between 2016 and 2020. So in the first quarter, we talked about an attach rate for our customers.

There are still -- the customers that are using AI products has moved to the teens, which is great progress over the last 2 or 3 years. We're in the high teens in terms of the number of people who have moved beyond RPA. There is still room for expansion for that base to grow.

The second piece is that we look at the world in terms of deterministic automation and probabilistic. So deterministic is rules-based. You want the same outcome every single time. Probabilistic automation is where we see agents, software with agency come into picture. So when you look at our strategy, fundamentally, we want to transform the way people work by creating end-to-end automation. That requires 3 things: Being able to do it on a rules-based, which is, like you said, kind of our founding technology in RPA; being able to have all the extensions around it, which is what we've done with the platform between 2018 and 2021, and now incorporating Agentic.

And then if I break down Agentic into 2 pieces, we are the Switzerland for Agentic. So there are companies out there that will do Agentic for CRM, Agentic for their HR applications. We will do it across applications. Many companies have 10,000, 15,000 applications in their tech stack. So in order to really automate, they need to go across.

The second piece is we're going to lead in governance and controls. If you're in health care doing claims processing, the governance and controls is a significant part of your decision making. And our first quarter -- our deal that we talked about in our first quarter earnings, that highlights where we're differentiated. When a Fortune 10 company picks us as their leading Agentic platform -- as their leading Agentic technology.

And then the last piece to underscore is we also launched Agentic Orchestration. Think about a world where you're sitting down and you have applications, but then you have robots doing transactions, agents doing transactions and humans doing transactions. You need 1 orchestration layer to be able to make sure that happens. That orchestration layer, that is something that we launched here in -- at April 30, incredible, incredible feedback from our customers. And that is part of our strategy to move not just at the transaction layer, but the overall orchestration layer as well.

That's great color. So if I remember during the strong adoption of RPA time. A lot of companies started their chief automation -- center of excellence kind of new role started, right? So now are you seeing that's the same kind of department extending becoming like Chief AI Officer, like -- or is it now -- what are you seeing -- the same customer...

Yes, it's different. I don't think any 1 customer has 1 approach to AI, which is reasonable and kind of understandable. So in some cases, the COE that was doing RPA is extending their scope and doing end-to-end automation, inclusive of Agentic and AI. In other examples, we see ahead of AI that has -- that we are selling into that is in-charge of the AI decisions and really driving the Agentic world.

Interesting enough, the few that we have started interacting with, they actually came from the RPA COE, which is very -- which is fascinating and obviously hopeful for us as well. And I think in other areas, it's spread across different areas. So you can have it under the CTO organization, making the decision, the CIO organization. I think all of them, depending on the company, has different governance models.

What I will tell you has not changed amongst all of that. ROI drives the decision because the lines of business are funding the next round of implementation. I think in the past, you saw all these innovation funds that we're moving and giving temporary burst of things.

ROI is really defining software evaluation today which I think it's a little bit economic, but I also think it's a little bit of the learning in the past 3 to 5 years of AI of a lot of overpromising and underdelivering from just the overall market. And frankly, when you look at ROI, that's where we feel like we have a really good advantage. We can provide really tangible ROI to our customers.

So what kind of adoption you're seeing right now with some of your, I think, Maestro and Agent Builder, some of the products already in the market?

Yes. We have hundreds of proof of concepts right now that are going there. So think about our 10,800 customer base, our top-tier customers, we're seeing enormous penetration of getting the proof of concepts moving. The #1 request that I'm getting right now is the need for more presales in technical engineering on the front end of our processes. To me, that is a great sign because people aren't content with a video or a PowerPoint, they want to get the pilots and the proof of concepts moving. So we see incredible demand from there.

We closed Agentic Government for the federal government. That is helping operations officers using Agentic technology, freeing their time, whether it's logging -- whether it's logging weapons or different inventory -- inventoriable items, that is something that we are able to help with.

The health care deal that we closed in the first quarter -- that we closed early in the second quarter that we talked about. Sonic Automotive is a third one. So we do see tremendous adoption. And we see hundreds of thousands of agent runs for a lot of our private preview customers which means they are truly putting their hands on the technology and running them in production and getting to understand what their impact could be.

So this is the question I always get from investors is Agentic AI, I know it's pretty early. But when you look at the opportunity, you heard RPA a certain opportunity to automate some kind of workflow some actions. Now when you think of AI coming in, is it going to replace part of it? Or is it going to complement or augment it or is partly cannibalistic? How do you see that opportunity now?

Yes. It's -- from my standpoint, it's complementary and augmentative. I can't say there's not going to be any cannibalization. I think that would be disingenuous, but that's not the primary sake. And I'll just give kind of like 2 reasons why. You remember, in personal productivity, you're not as worried about rules-based automation versus probabilistic.

If I say download 10 e-mails or download 10 attachments, there's very little chance for error. If you look at our customer base, health care, financial services, public sector, you are not going to take a chance with a patient record of their medical history to be sent in a with a probabilistic piece of software.

It is -- there's a risk element to it. There's a cost element to it. And even if cost comes down, there is a complexity of maintenance to it that you'd have to contend with, right? So in our view, when you look at a workflow today, you're going to find a lot of agents that are -- agentic opportunities that are hovering on the next step after a robot where a human would have gone in. I think that's the first.

Just one interesting anecdote, and I've said this in a number of my one-on-one tables, we had a partner that did 20 customer roundtables, in those roundtables of 100 ideas, 70 that came out were actually robotic ideas. Customers don't know the difference really nor should they between probabilistic and deterministic but their technology arms do and they do not want to put probably -- deterministic processes in a probabilistic framework for both what we talked about is, complexity and risk. And so I look at it as augmentive as you look for agents, I think you're going to find robots. And if you look around your robot, you're going to find agents.

Okay. So in terms of adoption, I think, if I remember pretty strong attach rate to the $1 million plus ARR I think, 85% or so, but that's still 20% of the base. So what could you do further incentivize that kind of adoption?

For the $1 million-plus base, Siti?

Yes.

Yes. I think 1 is just being close to them. I don't think we have to incentivize. I think the ROI is going to incentivize that. The main thing is education. So I think in this world of like AI confusion, which ebbs and flows, we have to make sure people really understand what software is capable of doing. Just to give you an example, we don't have to produce any of our own LLM models. We partner with everybody, and we have very open architecture.

Sometimes we get asked, are you tied to any 1 model because that creates security restrictions, or vendor restrictions from their standpoint. So I think there's a lot of education that goes into the first round. So from other incentives, I think it's really the ROI that will incentivize them. I would also say getting beyond the top base and getting to the rest of the customers, that's why we're super excited by the partnerships with Deloitte and our regional partnerships for smaller customers like Lydonia, TQA, Accelerate, and we've become very close with them over the last year as we've simplified our partner program focused on enabling them and really invested in them to grow and drive adoption across our customer base.

The other topic I want to hit on is your recent acquisition of PEAK. That kind of get into more of a vertical strategy, that was kind of a surprise in the first place when I saw that but you talked about the vertical strategy there. Can you help us understand how you're thinking about? Is it more -- doing more tuck-in acquisition? Maybe talk about your strategy itself.

So I think horizontal is a strength first. So just to give you an interesting metric, of the customers that we did -- that we have in our private preview, there's no use case dominated more than 10% of the number of use cases like not one, which means that a horizontal platform unlocks the ability for a customer to go after where they see ROI, and it also speaks to the ubiquity of what our platform can deliver.

The area where verticalization comes in, horizontal is great because it has a massive TAM. At the same time, it is a harder sale in some cases, because you are reliant on the customer to figure out what they want to go and solve.

Verticalization has -- will have slivers of that TAM, but it will cut down the time to value by giving them software that immediately hits the outcomes that we're looking at. So if you look at PEAK, inventory optimization and pricing optimization, gathering all of the data and being able to do the transaction in the system to make decisions on whether you're producing inventory and if I add prices, what does that do to sales and then correspondingly to production. That is a great outcome that is relevant to any customer that is -- has high number of SKUs as a part of their business model.

For us, we've seen great uptick, great interest right when we brought a PEAK into our company. So I think we can selectively go after verticalization to drive faster time to value while still maintaining the horizontal benefits that we have of our overall platform.

And this AI agent -- how do you think the pricing would be? Is it more users based, agent-based, seat-based?

Yes. Right now, we've rolled it out in what I'd call more of a pseudo consumption model. So we -- similar to MACC credits, that you get on the Microsoft side. People will buy bundles of use it or lose it, but it will be used upon consumption. So we can monetize a bundle of x, or 10,000 units. And every time they do an agent run every time they consume a model, it will draw down against it. That's our current pricing model that we're going with.

Last year when you were here, you talked about the go-to-market transformation. So I think recently, you talked about that I think you completed like how does the now current go-to-market strategy, position you for the growth or the next phase?

I think our go-to-market has never been stronger and more stable. So if you look at the average tenure of our leaders, there are greater than 2 years of UiPath experience now that are sitting in seat. The average tenure of our field has increased because we've moved the organization less and the number of accounts that have been switching hands because of various initiatives has come way down.

We're getting higher -- better customer intimacy, more stable and probably the most impactful thing from my perspective, we eliminated all of the central organizations that really put distance between us, our product teams, our management teams and our field teams, which gives us -- which has made us faster and more in touch with what customers are happening.

If you look at Agentic, Our ability to respond to customers fast, if you look at our Agentic launch, making sure there is consistency of message and consistency of approach. Both of those are -- all of the changes we made I just don't think we would have had this much momentum with our Agentic launch. Had we not made those changes.

Okay. I want to ask you on the AI side that I've been asking other companies as well. How are you using AI internally for your soft -- product development, go-to-market any of the areas?

So we've definitely employed AI across our engineering teams. So a lot of the code is generated by AI internally, and we're looking to increase that.

The second piece is we are deploying Agentic. We are customer-zero, we have been. So just yesterday, I did a meeting just on the number of agents and Agentic ideas that are there. Similar to our customers, we are literally starting with all the workflows that we automated with the hundreds, if not thousands of robots and looking around and saying, what are the Agentic ideas behind that. And within 3 weeks, we've come up with hundreds of opportunities to be able to Agentify UiPath using our own technology.

The third piece is we are -- we employ an experiment with tools for personal productivity. Whether that is Anthropic, whether that is ChatGPT, we're going to go and experiment and see how it is. It's been -- honestly, the success of all 3 of them have been awesome.

I was talking about it today, like we had 1 of our groups, they were able to take 4 hours' worth of work to condense and understand and help the documentation for a cloud migration for a customer between our platform and different reasoning models, they were able to bring that down in 2 minutes versus the hours that it would have taken, so super excited about what's -- what the potential is.

Now you're having a dual hat of operating role and CFO. Do you see this is kind of 1 of the key driver of operating leverage, not only your company for other companies.

100%, 100%. I think you get to choose whether you want to reinvest. But is there another step function change in productivity? I think you would have to be in a significant minority to say, no.

I think the biggest thing is having the leadership culture and incentives because it's deeply personal to people, right? When you do just -- I automated a forecast for ourselves using 2 reasoning models. It's hard to admit that, that reasoning model did it faster and more accurate than the hours that an analyst may put into it. So it comes with a certain sense of humility and ability to transform. But 100%, there's a step function change, not just for UiPath, but to me, the overall industry and our entire customer base.

I mean your comment certainly validates our -- 1 of our -- my team, our team published a note. So it's a plug, but we did it internal AI, how a company using and bring operating leverage, we quantify that, yes.

Yes. Interesting, all of the analyst research reports that come up after earnings. Just to give you an idea of deterministic and probabilistic, we were able to, and Jake sitting in the audience here, we were able to go and gather all of that, putting that into a reasoning model and get a quick summary of 36 -- or whatever is the number of reports and number of pages, and it came out in minutes.

Now the interesting thing is you still have to go and gather all of those documents, now attach that with a robot. And now attach it with an agent to say, bring me all press releases as well that you think are relevant. You're giving it a probabilistic task with a deterministic task of gathering all of the documents that you specifically want. So the agent doesn't overlook something. And then you use a reasoning model to summarize it and provide an output. That is a simple workflow that you can put and watch it happen with UiPath Maestro.

Let me pause here and see any question in the audience. You can raise hand.

Do you see any competition from the traditional BPO companies? Or you are a totally a different company?

It's an interesting question. They are some of our best customers. So just to put that in context and some of our best partners. At the same time, I think there's always going to be a competition between labor and technology, right?

So we're -- we definitely don't play in the labor. We look to augment labor from that standpoint. I think different companies take different tasks. I think the BPO is about timing. I don't think it's about outcome. So outcome 1 is a customer automates and then takes whatever they can automate and moves it to a BPO.

Outcome 2 -- another -- the other route is, a company moves it to a BPO and the BPO uses technology and then their own skill sets to drive efficiency, right? Either way, technology is employed. It's just a question of when. So in my mind, I don't see them as a direct competitor. I look at them as really good partners, really good customers. At the same time, they have -- they do introduce different paths to get to an end outcome of technology driving efficiency for a company.

So in the same context, your traditional competitors when RPA, how has that evolved the competitive landscape as you're looking to the next phase of your Agentic AI automation?

I personally feel like they've been left behind. Like I look at our competitors as we are now going to emerge as competitors to other companies, whether it's Salesforce or ServiceNow or different things for different workforces, but I think we've leapfrogged those competitors, both numerically as well as technologically.

And to be candid, we leapfrogged them technologically, which is why we've leapfrogged them numerically. So in my mind, while we will see them in pure RPA at small scale, we -- our win rate is very, very high. We don't really lose customers. We do see areas where we can displace them. But the Agentic landscape, that's a different group of competitors from my standpoint. That's not -- we're not playing against them.

Okay. And probably last question here. We talked about some of AI agent, agentic solution. When do you think that's going to uplift your revenue growth? Is there something in this year, next year?

So I think this year is around proof of concepts. So I think when you look at POCs and pilots, to me, they are going to validate the technology and they are going to validate the ROI. So I think in the last couple of years, a lot of companies have been burned by the overpromise of AI. And so ROI is going to be the defining characteristic for revenue. So I look at it this year being proof of concepts and pilots, and I look at -- as we get beyond this year, we definitely see there -- we definitely see meaningful revenue potential, coming from these -- from this technology in this market.

I look forward to hosting you again around this time, and we'll talk about more use cases then.

Thanks so much, Siti.

Thank you. Thank you, everyone.

Thanks, everyone.

Okay. Good morning. Good afternoon, everybody. It's Keith Bachman here from BMO. We're part of our ongoing virtual software conference. We're thrilled to have a shine from UiPath and so with that normal process here, we're going to be about 35, 40 minutes. I'm going to ask questions. I think there's a way that you can also pose questions or candidly, just e-mail me directly, and we'll do our best to get to them.

So with that said, let's go ahead and get started, Ashim, again, on behalf of Bank of Montreal, thanks very much for joining us much. Appreciate it.

Let's talk about -- a little bit about -- we're going to do some big picture questions. And then without much surprise, we're going to jump right into AI. Let's go to the big picture, though. You recently reported your quarters, we think about calendar year '25, how do you think about, a, the puts and takes this year on net new ARR and a way to ask the question within context is also how do you think about, say, the coverage ratio of that guidance, meaning the pipeline, the visibility versus the years past.

Awesome. First, thanks for having me, and thanks for everybody for joining in. So look, as we look this year, as we typically did for first quarter, really looked at the environment, we go through our FP&p model, we go through AI models, data science and also go through a lot of customer and customer discussions. I would say second quarter is relatively consistent to what we saw what we talked about in the first quarter. So from a guidance perspective, the guidance what we see in front of us. So we evaluated our pipeline in -- so as we're talking about our current forecast, et cetera, my feel is very, very well covered because it is we guide to what's us. We continue to take a very prudent approach to our guidance. The macroeconomic environment variable. I think you can turn on your I don't think that's much in my mind. There's hopefully here of some solid we talk about our customers, whether it's policy, whether it is an economic or geopolitical, I think there just remains a lot of -- there are more questions than answers, frankly. And you look at that always as having a weighing on the environment similar to what we talked about in the first quarter.

The second is the federal transitioning -- transition in the U.S. It's still evolving. And we remain prudent on that. So when you look at certain agencies, we're still waiting for full confirmations in certain places of some key roles. There are -- for all intents and purposes, there continues to be DOGE deep dives and reviews.

We are super excited about where we're positioned with the federal government. But just given where it is in the transaction, we do remain prudent around that.

The Agentic launch, we feel very good about it. The -- the proof of concepts, the momentum that we are seeing early from our customers it is very encouraging, both in terms of the product feedback as well as, and I know we're going to talk about this, the synergy between our workflow automation, or RPA platform that we built between over the last 20 years or Daniel has built over the last 20 years, as well as the agentic offering and agenetic orchestration.

I think everybody is -- this year is around proof of concept. So it's early innings of a large and growing market opportunity. So we're -- we haven't factored in major revenue upside from Agentic at this moment, which we think is the right thing to do. But overall, when we look at what's there, we feel good about our coverage ratios. We feel good about the signals from our customers.

I would tell you, I don't think we've ever been as connected as a team and as connected with our customers as we are today versus the last 3 years. And that's a function of the restructuring, leaning out a lot of middle functions that have existed within the company.

Well, let's pull in a couple of those threads. One, I want to just clarify, given what you said at the outset about your macro -- or excuse me, your coverage ratios, because of the Fed and perhaps some of the other issues, would you say there's a higher level of conservatism or scrutiny this year? Or is it the same process where you put your filter on and you come out with your forecast? Is it the same in the last couple of years or it's a tweak higher?

My view is we're better. And the reason for that deal is not because of the federal uncertainty, just more again because of the connectivity we have with the team. I don't think that means that we were bad in years past. But I just think that there -- when you have -- if I separate you and maybe with 5 people who are all talking, it's harder for you and I to have signals together. I think we are much closer to our customers, much closer to our deals than in the past. And I think that is structural and cultural as Daniel stepped back into the CEO role over the last year.

But otherwise, I also think we're taking a lot of the good things in the past. The data really being understanding our FP&A models or data science or coverage ratios, I think we have a very good understanding, and we've remained consistent on that as well.

So you think your visibility and your pipe is higher as a way to rephrase it because you just have better data?

And also better customer connectivity. I think that's very important,stronger customer connectivity.

Yes. I hope you don't mind, I'm taking notes while I'm talking to you. So we're doing 2 things at once, which is usually a little dangerous. As you think about -- just to sort of ask it more bluntly, you said our organization is better.

Path has gone through some changes since the IPO and from a leadership perspective, some organizational and more so than other companies that I've had visibility to post public. And the 2 parts to the question are why has there been more sort of, I'd say, organizational changes; and b, is everything sorted out?

Yes. So I think the first one is a hard question to answer. And I don't -- I've seen companies that have a significant change when we put the comparison. I think one is the scale at which -- the process of -- the speed at which we scaled and went to the IPO, I think was also unprecedented.

So I think when you get into that significant scale at a, what I would say, almost a record pace, I think you go through different structures to understand what really works for the company. And so I think I have no regrets about the past.

I think Daniel has talked about his reasons for stepping back into the role. And I think a founder-led company has been -- has really restored us to our roots.

What I would say is the changes of the past actually solidified my conviction that we do feel like we have the answers today. And those answers rest not solely on people, but on principles.

Principle one, less central organizations. Central organizations create a lot of churn, they create a lot of cost. And I think when you look at a lot of the leaders that have come in and out, there's also a lot of roles that are within that -- that have been more central or internally focused than external.

The second principle is customer first. And I think that lends itself to the first piece.

And the third is really around speed and innovation. And you see that manifest itself just with the speed about which we brought our Agentic platform into the market. So I think if you take those 3 principles, I feel very good about where we are. Stability is a big point of emphasis for us this year.

When you look at our sales leadership, from the outside, it feels like there's a lot of changes. Our key sales leaders have more than 2 years of experience with UiPath. One, our Americas leader kind of had left and then came back just given the changes we were making, and we're super excited to return to the company. When you look at our major vertical leaders, they've been in seat for multiple years now. Our average tenure of our sales force continues to improve. So we feel really good about the stability that we are driving within the company from our leadership team down.

Okay. Okay. Let's -- I want to transition to the Fed for a second. With all the changes in Elon leaving and what have you or -- do you feel like the scrutiny is the Fed or dose has tweaked down a little bit? Or is it still the same pressure?

I talked to 3 federal customers just over the last 15 days. I don't think it's about Elon leaving or DOGE screening. I think there's just tremendous pressure given the ambition and the transition of the current administration and the goals that have been set. Current -- and so I don't think that pressure has been gone.

What I do see as improving is there are starting to get more clarity around who are the decision-makers within the administration. And I think that will enable budgets to being set and priorities to being set. What we're excited about is, as we talk to a lot of our federal customers and federal partners, Department of Government Efficiency aligns very well with what UiPath does. People able to do more with less, that is very aligned to where we are positioned.

And so -- we are really pleased with the renewal rates we've seen. We had the Agentic Airman deal that I thought was a great deal in the first quarter that closed that shows and highlights a combination of like how deep we are with some of our agencies in the Department of Defense is that example. As well as the promise of Agentic.

So we feel very good about our position, but we have to be patient to allow the administration to kind of work through and the agencies work through the changes that they're undergoing. And I think that's going to continue through the second quarter, and that's what we factored into our guidance from a prudent standpoint.

And maybe just review how you guys are thinking about the Fed in terms of the ARR growth that you've provided.

Yes. We definitely saw -- we definitely accounted for first half seasonality to be impacted. And you see that in the numbers. We talked about that in March, we really didn't change much on that. .

We do see and continue to track a prudent level of activity for the second half. It improves, but it's still prudent from our standpoint. Just given the uncertainty that's there, we're assuming that there's going to be a level of transition that continues throughout the year.

But second half seasonality is a little better than the past in terms of net...

Yes. I would say overall volume, we feel like has come down in the Fed, just given the uncertainty for the year. And the seasonality is more back half loaded than front half loaded. So we feel like the back half is you start seeing the moratoriums release the budgets kind of getting solidified. And we're very -- while we've taken a prudent approach on that, we definitely see some level of rebound of the federal business in the second half.

Okay. Okay. Let's transition to products, and we'll just jump right into agents and Agentic offerings. And this question you and I spoke about after your report, but I want to revisit it. I'm one of the few people on the sell side that actually covers a small company like IBM as well as ServiceNow and the way we're -- we're seeing the world is you have the application vendors like Salesforce and Workday and others offering their agents within the scope of their portfolio, then you have horizontal plays, including agent orchestration for companies like IBM as well as ServiceNow.

And so I'm just trying to figure out maybe review for us or revisit why you think you have a right to win in this offering what is increasingly a crowded field associated with Agentic capabilities and Orchestration more specifically?

Yes. The first thing is, I think crowded also defined by the opportunity. And I think we don't think there is a one winner in this space. It's too large of a space.

If you look at every manual task that is being done, that can be automated with rules-based automation and then further enhanced by reasoning or probabilistic automation, we feel like it is a massive, massive market. And so we are one of the players that has a right to win, and I'll expand on that here as we go through the question.

The first piece is to really understand like what's UiPath platform fully. We have both our traditional workflow automation capabilities, RPA, document processing, communications mining process discovery.

I say this, and it doesn't always move the needle from a center. We are not an RPA company. A year ago, even pre-agentic. It was not RPA. It is the end-to-end automation platform that existed. And now when you add Agentic capabilities, it further broadens that and then you put Agentic Orchestration, which allows us to orchestrate agents, humans and robots you really don't find a company that has that breadth of automation capabilities as us.

The second piece of it is, there is an inherent link between Agentic Orchestration -- Agentic Automation and deterministic or robotic process automation. There you are -- the notion that says that all automation will be probabilistic does not make sense from a government standpoint, from a cost standpoint and even if costs came down, which we do see from a complexity stent.

The third piece is we're Switzerland. So I think Salesforce has incredible platform, they are going to be more geared towards the Agentic Automation within their spaces, right? You see that for companies that specialize. We really are the Switzerland. So whether you are a mainframe application within a large bank, whether you are a newer company, we really sit across the entire stack and automate across it, which gives us a difference.

Now that's the vendor talking. I admit, I'm representing the vendor. When you look at our customers, Agentic Airman closed in the first quarter. When you look at kind of second quarter, we talked about closing one of the large deals or a significant inventive deal with a Fortune 20 health care company. And they really -- we had to sell into the head of AI and they look at real transformation, both with our RPA platform, but really with the Agentic capabilities and in Agentic Orchestration that we're doing.

We have hundreds of POCs right now that we feel really super positive about. And customers are seeing that if that ability to do so. And we are also verticalizing where appropriate. If you look at our peak acquisition, we are now agents that can do pricing and inventory optimization. We are specializing agents within health care and within financial services.

So our customers have really been pleased with what they have seen. And just to give you a couple of examples, a medical device company, they're implementing agents into their revenue cycle management process. Now why that's important? When you think about the best place to find agent opportunities, go to find where the robot and human are working together, because there are things that, that human is doing that now an agent can do.

So when you have 10,800 customers, hundreds of thousands of workflows that are in production, you really can go to that set of workflows and start looking for Agentic opportunities. and we feel really comfortable about that also being a core differentiator for us and giving us a head start kind of in this race for customer trust.

All right. Right.

And sorry, I did forget one last thing. As investors look across, I really want to emphasize, there's personal productivity, summarizing an e-mail.

Yes.

Here is enterprise-grade productivity. The level of governance and controls that are needed that are inherent to our platform to automate mortgage processing. Patient record management, revenue cycle management, customs, that is significant, and that is inherently differentiated within our platform as well.

Okay. Before we move off the competitive context, I just wanted to hear your thoughts on Microsoft, which admittedly, in my opinion, is behind an agents. They had sort of a toe in the water in RPA. It seems like it's not a focus for them. But how are you guys thinking about Microsoft as you articulate your strategy on humans plus robots plus agents?

So one is, I think we have a really good product partnership with. Microsoft ironically internally is more of a friend than it is a foe. We have really good relations all the way up throughout the entire Microsoft senior management.

Our Azure consumption is actually significant and is important too. So they are interested in investing in bringing workloads to our platform because that, in turn, creates workloads for Azure. So they are a partner, and that's why even 3, 4 years ago, even when in the RPA space, forget about advanced Agentic capabilities, they talked about us as being their preferred automation partner and vendor, that's there.

The second piece is, I think when you look at Microsoft, they still are going to be geared more towards that personal productivity side. We differentiate really through the enterprise-grade productivity that we want to go after. And we look at that as separate spaces. And frankly, the feedback from many of our customers is those are separate capabilities that are there. And that's why you continue to see net new ARR growing for us as a company despite Microsoft now having been in the "automation space" for the last how many years since IPO, for us.

Right? Okay. Let's talk a little bit about the cadence. Break it down into 2 parts, and we're staying on Agentic capabilities for the moment. You mentioned you have hundreds of POCs, how are those POCs unfolding? What I mean by that is, are these more experiments within the context of companies?

And sort of the root of my question is big opportunities you identified, why is it not moving faster? I guess you alluded to it in terms of governance is probably one of the key sources of friction. I know it is for Salesforce and others. But maybe just talk about kind of the pipeline of Agentic opportunities within those POCs and what's the source of friction?

Well, one is, I don't -- I kind of don't agree that it's not moving fast. I think it's a question of expectations and perspective. So first is we just launched our platform at the end of April, right?

So from that perspective, we had already closed -- we have a number of customers in private preview. We had already closed some of those deals that we've talked about, whether it's Agentic Airman or whether it is the health care example that I gave.

The second is, I think it's unrealistic in the enterprise productivity in the enterprise workflow space to say everything is going to kind of like just shoot though in a couple of months. And in fact, a lot of customers had been burned by overpromising on the AI hype, and we see that eternally contributing to making sure there is more proof of concepts that have been really tried, true and vetted as they're making their decisions and their investment decisions, right? And that is something, I think, that's been well covered, and that's been well heard by us from our customers.

We actually thrive when it comes to a bake-off because we have product trade. So if somebody comes and says we want to do a proof of concept, it actually makes us super happy to do that. in terms of where that looks.

The third piece is, I think when you look at -- I came from a large company in GE. As fast as companies change their organizations also need time to adjust to it, right? So you have the entire organizational dynamics and security dynamics and all of these things that are taking shape within an enterprise. And I think those things just take time.

At the same time, having hundreds of POCs right after you go into your GA, the activity that we're seeing with our customers is super exciting. And frankly, the success rate of our proof of concept is also super exciting. So you look at those things, and I feel like we're moving at a good speed that's realistic. And it's become cliche, but everybody says, everybody overestimates the short term. and underestimates the mid term, I think we're trying to be realistic about that.

So if we look at this year, is really proof of concepts, but we see this as a meaningful opportunity. It should be bigger than the RPA opportunity. And the RPA can help scale us in 5 years from almost nothing to $1 billion.

Right. If you think about -- I wanted to ask the second part of the question is you said a little bit of contribution this year, but more meaningful contribution in CY '26, right? It's been the Agentic kind of backdrop and is there enough to make some contribution in Q4 though as in net new ARR, but it's really or we should be really focused on CY '26?

I would be more focused on CY '26. I've given my guidance. I think I factor that in, we've talked about nothing meaningful for this year. Nothing meaningful doesn't mean nothing. I think that's really important. I think sometimes people forget that word meaningful in that discussion. .

So in our minds, we really look at this as the year of proof of concepts. The first set of customers coming on board as you get those success stories. I think success breeds success. So a successful '25 in terms of fiscal '26, I think will breed a successful '27 from that standpoint.

When you're in the POCs, presumably, these are done RFPs or there's competitive dynamics surrounding even doing POCs, I would assume. A, is that true? And B, who is also in there with you trying to -- to win those workflows?

Great question. It's interesting. You don't get a lot of RFPs or RFQs in a lot of examples, like as I talked about that 10,000 customer installed base, we are getting called in just because people are like, wow, we know UiPath, we've got great success with UiPath. They hear our launch, they know our products, our reps and our customer success teams and presales teams are very close to them, and they just want to see the product in motion.

In the examples where you see kind of real tests and trials, that are more competitive, actually internal, people are testing their internal capabilities. Is this something we can build? Or do we need [indiscernible]? And what's fascinating is, I think the models people think that they can build.

The platform, the governance, the orchestration, the connectivity, that has really been able to differentiate. So we didn't really talk too much about Maestro. Maestro connects agents, robots and people. It delivers fixed workflows at the enterprise level, and it gives you insights into those workflows. That built on a modern stack is a real differentiator for us, and it's almost impossible for an internal organization to replicate that, and that really gives us a step up versus just saying, how are you incorporating the model into an automation or a workflow. There are multiple tools and there's multiple dimensions that we are able to win in that discussion versus internal.

Okay. Thank you. Maestro is going to be my next question. So I appreciate you bringing it up. Ashim, how about the journey from POCs to revenues, again, not this year, but what does that process look like? In terms of this genetic world, combined with Maestro, the whole category, how should we be thinking -- how does that translate into -- I shouldn't have said revenues, is more think than ARR, but what does that journey look like?

Look, I think proof of concept shows proof of -- there's 2 things that you've proven out, technology and ROI -- so I think the proof of the technology gives you the right to play, the right to compete. I think the proof of the ROI gives you the right to revenue, the right to ARR.

And so we're not going after our proof of concepts of a few demos or PowerPoints or precast videos. Our proof of concepts are saying, "Can we make things faster, better, more accurate?" And in my mind, via, as we solidify that, you're going to start seeing like not far behind proof of ROI is realization of revenue.

And then the question becomes like what is going to be the bite, the first bite? And how big are the subsequent bites from it? And I think that is going to be a function of both the environment as well as budgets as well as kind of internal organizational culture. And I don't have a formula for that. I think it's different depending on the technology.

If you look at the early wins, we're really excited by the mix of some companies are going all in. both in their proof of concepts as well as organizational. But I think that's going to be a relatively similar mix to the past, but we're going to have to see how that unfolds.

Okay. Ashim, one is -- as you're, I'm sure, well aware, and Jake also. I think the reason UiPath has what we would characterize as a relatively modest multiples is because it's you to be still an RPA company, which you've asserted as an inappropriate reflection of the business. Part of the way to respond to that is to give more granularity about the business mix. And so would you envision at some point in time, giving metrics associated with legacy or robots versus other? Or do you view them now as all intertwined, so that's not appropriate?

Well, first is, I'd point everybody to kind of our fourth quarter earnings back in March. We did disclose RPA and the attach rate of our AI-based products. So when you look at that, I think we said 18%, if I recall, but we can get the exact number that was there.

But when you look at -- the first thing is we have increased our disclosure around it. The second, as we go through it, we always want to -- we want to give as much granularity as possible. We increased our -- as an example, the disclosure around customer cohorts. So the customer is greater than $30,000 and how they grow.

So we will always look to give as much disclosure to give the best investor decisions. It's a double-edged sword. So once you give it be, obviously, then we committed to continuing to give it. And I think there's a maturity that we also want to make sure is realize that's there. At the same time, I think Agentic, we're giving a lot of transparency into the script of our Agentic wins around -- about metrics around our trials, the number of customers, the our agent runs that we've seen. Formal revenue disclosures will continue to evaluate.

Okay. Perfect. Okay. We only have about 8 minutes left, and I've only gotten to 2 subjects, so shame on me. With that said, if I step back for a second, I understand that net retention rate is a backward-looking metric. But with all these really interesting opportunities unfolding surrounding the expansion of your portfolio. If you think about over the next 12 months to 18 months, how would you envision NRR moving around?

Yes. Look, I think first is it is backwards metric. It's a pre -- into trailing 12 months. We've talked about this year being a point of stabilization. And then if you adjust for macro and Fed, that is essentially what we feel from our internal execution and marketing capabilities.

So I look at like when you think about our second half guidance, you can see implicitly what's assumed in there. We don't see a substantial shift from new logos to -- from expansion to new logos. So it's pretty easy to see the trajectory of net new ARR, that's there. And again, that's with no meaningful Agentic revenue inside implicit within it. So I feel -- I honestly feel good about what the future holds and the stabilization of not just a metric but overall metrics that we have.

Okay. Yes, we actually do that breakout. Let's shift gears for a second to a small company. And how do you think about the evolution of your relationship with SAP? And how does your Agentic maturation with your product portfolio, how does that impact the relationship with SAP incremental opportunities?

Look, I think our partnership has really started with kind of our core automation portfolio because we had just launched Agentic recently. We'll continue to explore not just with SAP, but with every company, the ability to partner with us. We look at partnerships as a strategic lever for both mindshare as well as wallet share within a company. .

That being said, I think we want to -- we still see tremendous opportunity with the S/4 HANA migrations for getting our core automation platform out there. Just to put it in perspective, we ourselves had implemented our ERP here internally, and we had a 93% clean core.

Just if you kind of think about that, that is a step function and a game function change for ERP implementations, and that is really possible by our technology and UiPath platform. I think SAP is the same thing, and that's why we're continuing to focus on driving joint customer pursuits together. And that partnership, like you said, it's a pretty -- as you indicated, it's a large company. It takes time, but we feel positive still about the overall potential for what SAP could bring.

Okay. And Ashim, I know you're an automation company and you sort of have the business that you grew up with in robots and you've broadened the portfolio quite broadly. But do you feel like the legacy business, for lack of a better work is at a point of stabilization? And some of that you've had, what I'll call, low end customers that have been degrading a little bit. That's probably more of a customer issue than an ARR issue. But how should investors think about if they want to break the business into 2 pieces, the RPA business unfolding versus all these new opportunities, Maestro and Agents.

Well, you can -- the first thing is, I really don't -- I will want to change the concept of the legacy. I think it's still very relevant and present. So why? Just think about the number of tasks that still remain to be automated across enterprises. It is a highly relevant technology. The piece of technology has moved so our attention span seems to change. But actually, it's still -- from that perspective, we're still -- in our minds, there's still a large and growing market.

You can look at IDC's numbers of double-digit growth within that segment or in that area of automation, excluding Agentic. And I think that, that is the first piece that I would emphasize. So if I think about foundational or deterministic automation, we still see that as a growing base, we see that as attracting new customers. We see people who are still able to expand across departments, to cross-sell into different capabilities within workflow automation.

To break apart Agentic and RPA is like breaking a part of workflow. So in our minds, it is highly synergistic to have robots, humans and agents working together. And so from our perspective, when you go to the robotic workflows, they should lend itself to more Agentic workflows. Just go to any workflow where a robot is working with a human and say, what is that human doing, there's probably an Agentic capability that's there that can be unlocked in that existing workflow.

And let's say we unlock a new process, and one of our partners actually said this. They did a workshop with a number of customers. They said one of our conclusions, which is interesting, every idea that in the first 2 days that started actually was more of a robot idea or a deterministic idea than a probabilistic.

Well, I sort of asked the question in that spirit in that, do you think customers understand the difference in terms of or is that actually creating confusion on do I use an agent or do I use a robot?

I don't think it creates confusion for the right people who have to make that decision. If I was a business user, my goal is to say, how do you make my life easier. And the fact that UiPath has both robots and agents means I don't have to ask, do I talk to UiPath or do I not? We can do even [indiscernible].

Our goal is to have faster, more automated, less costly, more accurate workflows. When you turn it around, we feel really good about automation leaders, heads of AI, who know to say, if something is rules based, I am not going to go and put a probabilistic agent in there and create a 5% or a 3% additional chance [indiscernible].

Yes.

And I want something that has both UI automation, API automation and advanced AI capabilities that can go across any application in my stack. From that perspective, we feel really good about where we're positioned. .

Okay. Yes. Well, there's a lot more we can do there. Maybe I'll just ask my last question. I'm going to unfortunately have to condense it a little bit in the interest of time. It's just on go-to-market. In terms of -- how do you feel about the -- the stability and where you are now in terms of feet on the street as well as pricing -- the unified pricing models, what's the feedback? Do you feel like there's still some tweaks you may need to make? Any comments there?

Yes. I think first is go-to-market is stable, more connected with our customer. We're talking more and we're innovating more with our customers. So we feel really good about the core tenets of our go-to-market organization. .

Our partner ecosystem, they're a good measure for us. I feel like even in that area, we're more connected, whether it's companies like Lydonia, Accelerate, WonderBotz, which are regional or companies like Deloitte, where we're advancing Agentic ERP together. And EY, PwC, Accenture, we feel really good about our connection and our position there.

So I feel like go-to-market as a whole, we feel very good. The discussion around unified pricing, the first set of feedback has been really good. It's early, though. So I don't want to oversell that in the beginning. I think people like the idea of a simplified pricing model.

I think we and our customers book value the platform, which in our pricing 2.0 is more explicit on the value of the platform. And it gives -- it empowers customers really with more choice and more opportunity. So initial feedback is good, but I want to see how this year progresses before I declare victory or we declare victory in that [indiscernible] .

Yes, good practice, good policy. -- in, again, we're going to leave it there. We're 1 minute over, but I appreciate you coming on and spending time with us really interesting time for UI path, and we wish you all the best of luck, and we'll be focused on the journey and how you guys unfold all these opportunities in front of you. Again, thanks very much for your time. We appreciate it.

Thank you.