CS Disco Inc Aktie

Ladies and gentlemen, thank you for standing by, and welcome to CS Disco's Second Quarter of Fiscal Year 2026 Conference Call. [Operator Instructions] I would now like to hand the conference over to your first speaker today, Head of Investor Relations, Aleksey Lakchakov. Please go ahead.

Good morning, and thank you for joining us on today's conference call to discuss the financial results for DISCO's second quarter of fiscal year 2026. With me on today's call are Eric Friedrichsen, DISCO's Chief Executive Officer; Aaron Barfoot, DISCO's Chief Financial Officer; and Richard Crum, DISCO's Chief Product Technology and Strategy Officer. Today's call will include forward-looking statements within the meaning of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995, including, but not limited to, statements regarding our financial outlook and future performance, our future capital expenditures, market opportunity, market position, product and go-to-market strategies and growth opportunities and the benefits of our product offerings and developments in the legal technology industry.

In addition to our prepared remarks, our earnings press release, SEC filings and a replay of today's call can be found in our Investor Relations website at ir.csdisco.com. Forward-looking statements represent our management's beliefs and assumptions only as of the date made. Information on factors that could affect the company's financial results is included in its filings with the SEC from time to time, including the section titled Risk Factors in the company's annual report on Form 10-K for the year ended December 31, 2025, and the company's quarterly report on Form 10-Q for the quarter ended June 30, 2026.

In addition, during today's call, we will discuss non-GAAP financial measures. These non-GAAP financial measures are in addition to and not a substitute for or superior to measures of financial performance prepared in accordance with GAAP. Reconciliations between GAAP and non-GAAP financial measures and a discussion of the limitations of using non-GAAP measures versus their closest GAAP equivalent is available in our earnings release. And with that, I'd like to turn the call over to Eric.

Thanks, Aleksey. Good morning, everyone, and thank you for joining us. Disco delivered another strong performance in the second quarter as we continue to deepen our relationships with our largest customers, secure large and complex matters and extend our leadership in AI built specifically for litigators. We remain convinced that trust earned through deep litigation expertise, enterprise scale and security is what determines who wins as AI reshapes this industry and Q2 gave us further evidence that DISCO is earning that trust at scale.

In Q2, total revenue was $43.1 million, up 13% year-over-year, and software revenue was $36.8 million, also up 13% year-over-year. Services revenue was $6.3 million, up 18% year-over-year. Adjusted EBITDA was negative $3.4 million. We remain on track to be adjusted EBITDA positive in Q4 of this year. It was a great quarter, and I'll come back to the highlights in a moment. But I want to start with something we announced today that represents an incredibly exciting new era for DISCO, the launch of our Unified Litigation Solution. This will take DISCO well beyond Ediscovery and into serving the complete and unique needs of litigators with purpose-built AI capabilities.

In the near term, we're uniting the facts of a matter with the controlling law alongside proprietary litigation-specific workflows, all through a single pane of glass. We are ultimately building a full suite of integrated litigation solutions that span the life of a matter from filing to verdict, handling everything from legal facts to drafting to case strategy to trial preparation. While companies like Harvey and Legora have approached legal with broad general tools, we are laser-focused on litigators with a much deeper solution to meet the high bar set for scale, defensibility and security.

We've talked before about how litigators need purpose-built capabilities. How they want them in a single centralized interface and how no one has successfully connected the facts and the law in a seamless way that unlocks real strategic advantage. DISCO will change that. I'll let Richard get into the specifics a little later, but I want to tell you why we're excited about it at the highest level. It takes a litigator a decade or more to really master their craft. And even then, what one case team learns rarely carries over to the next matter or to the next team.

Moreover, the tools are generally only as good as the lawyer behind the keyboard. We believe a platform to understand the case and carries that context forward to put winning outcomes in reach for a lot more people and unlock a wealth of institutional knowledge and expertise that today sits siloed within individual lawyers and individual cases. This will mean greater ability to connect the dots, both within and across cases. It will give litigators greater control over the matters by assembling the tools and knowledge that they need to be successful in a single place.

Litigators will be able to focus on the things their clients value most, strategic advice built on deep insight. While DISCO takes care of the surrounding chaos with an integrated and secure system tuned to the things that are most important for the success of the matter. For litigators, this is a significantly better and faster way to work. For firms and corporations, this is a force multiplier that makes case teams more effective and more efficient with better access to information and applications built specifically for them. Our customers are already excited as we're developing this solution, we're working closely with some of our largest and longest tenured customers, some of whom are participating in our pilot.

One thing we hear time and time again during these pilots is that litigation is unique and the general purpose AI tools like Harvey and Legora provide only a fraction of the capabilities today's litigators need. This is particularly true when it comes to deep context, especially when it relates to facts and law. One leading national litigation boutique that is participating in our pilot program said combining our evidence directly with relevant case law is the missing piece for us. Integrating the two would be a total game changer.

Customer validation even at this early stage, gives us confidence that the market is ready for this and that customers see the long-term vision of what we're developing, building a solution for litigation intelligence and strategy unlocks a much more significant opportunity beyond Ediscovery as we believe that any innovative law firm that is serious about tapping into the power of AI to enhance the litigation practice will want the most advanced capabilities in their arsenal. We've been building toward this for a decade, Ediscovery, Cecilia, Auto Review and Advanced Research have all been steps along the way, and our Unified Litigation Solution is the next and most ambitious one.

As we're building for the future, the DISCO team continues to deliver today. In Q2, we saw strong performance across the 3 areas that matter most to our near- and medium-term growth strategy. First, continued growth in large matters. Second, continued growth in wallet share with large customers; and third, accelerating adoption of AI across our platform. Let's start with large matters because it's really the thread that connects everything else. We continue to see an increase in the size and complexity of the matters coming on to our platform. This quarter, we saw strong performance in both quantity of large matters and revenue generated from these matters.

As we've discussed before, larger matters are simply worth more to us over their lifetime. They generate more revenue, they expand as the case develops, and they stay on our platform longer. They also drive more AI adoption and usage on our platform. Alongside large matters, we also saw significant progress with large customers. The largest, most sophisticated firms not only have more large matters, they have significant litigation and frequent spending. So we believe growing this customer segment gives us a more efficient approach to a predictable and durable revenue base.

The number of customers generating more than $100,000 over the last 12 months grew to 354, representing $128 million or 77% of the last 12 months of total revenue. That equates to 15% year-on-year growth. We were also very pleased with customer adoption of our generative AI and agentic AI capabilities. This quarter, we saw continued growth in adoption such that revenue attributable to generative AI and agentic AI capabilities more than tripled year-on-year. Both Cecilia and Auto Review were drivers of growth in Q2 with Auto Review experiencing a material increase from Q1, driven by pipeline growth, larger matters, higher matter count and repeat usage.

I'm excited for the traction that we're seeing and for what's ahead. We and our customers view our AI-native capabilities as increasingly integral to how litigators actually run their matters rather than just a feature they tried once. The repeat use of Auto Review by customers is central to our strategy. In Q2, we began to roll out a significant enhancement to Auto Review that simplifies and accelerates the process of creating the inputs required for effective AI review.

As we discussed last quarter, our customers have varying degrees of AI readiness. This enhancement helps lessen the learning curve, making Auto Review an even better option for more customers and more matters. Advanced Research also continues to gain traction. The feedback from firms that have been piloting its capabilities has been unanimous in that it provides significantly deeper insights, reasoning and recommendations and is particularly suited to complex matters where its ability to autonomously reason across large intricate data sets adds another layer of intelligence to their workflows. We also saw very strong adoption of the DISCO platform, our new commercial model in the quarter, which continues to significantly outpace our internal goals.

In fact, we hit our December 2026 year-end run rate by June. The demand has been higher and adoption has been faster than we anticipated. We are seeing both more matters and larger matters coming on to DISCO due to our more straightforward pricing and because Cecilia AI is included on every matter. The apples-to-apples pricing and reduced user friction for our adjacent products has resonated with customers and further driven our performance. All of this adds up to an incredibly exciting story for DISCO. We continue to gain traction in our core business with larger matters, larger customers and accelerating adoption of our AI capabilities. This gives us a strong foundation to rapidly scale our next-generation Unified Litigation Solution. Now I'm going to turn it over to Richard, who will provide a deeper dive on those capabilities and the incredible work underway to bring them to life for our customers. Richard?

Thank you, Eric. As Eric said, the new DISCO Unified Litigation Solution is not a new product. It is the most important investment we're making as a company. So let me build on that and get into the specifics what the solution actually does today and where we're taking it. Let's start with today. Litigation is fundamentally different from the transactional and advisory work most legal AI has focused on. It's adversarial and winning hinges on the litigation's team ability to master the facts and the law. The facts that decide a case are often buried across millions of documents and not just the so-called hot docs that surfaced through the initial Ediscovery processes.

Winning strategy requires mastering the client's entire evidentiary record, leveraging all of the relevant governing law and reasoning across both together. And that is what this solution is built to do. The DISCO Unified Litigation Solution unites 2 things: your evidence and the law. It has native access to the Ediscovery database, not just the text of a document, but its metadata and the work product our customers have already built on top of it, and it's paired with DISCO's license to the full corpus of U.S. case law, court rules, statutes and regulations.

Concretely, that means a case team can get a live view of where a matter stands today the claims at issue, the elements that still need to be proven, the evidence gaps and the key dates all in one place. Underneath that view, the system reasons the way a litigator does. It maps each claim to the legal elements required to prove it and evaluates whether the evidence in the matter actually satisfies each one rather than just surfacing documents that might mention the right words. That's the difference between a general AI legal tool and a system with the native contextual intelligence to tell you whether your case holds up.

DISCO doesn't build technology for lawyers. We build solutions for litigators. The opportunity Eric described earlier putting better outcomes within reach for many more people really has 2 dimensions, and we're addressing both. Within a single case, it takes significant time for someone new joining a matter, a new associate, a partner stepping in to get up to speed because case knowledge tends to live in people's heads and scattered files rather than in a single system, anyone on the team can query. And the strategies and playbooks that were successful on prior matters mostly stayed with the people who worked it rather than compounding across the rest of the firm's docket.

Ultimately, our solution is designed to close both gaps surfacing what a case team needs now on a given matter, and, over time, helping the patterns from one case inform the next. Specifically, customers will create custom workflows, agents and tools that ensure their own best practices drive how our solution supports their work, both within and across matters. We believe this requires deep integration, not simple connectivity. And that's true for both the law and the evidence. General purpose AI tools and even other legal AI providers are largely reaching the law and the evidence through outside connections, an API call to a research service or a read-only link into someone else's Ediscovery platform.

Or even suggesting that lawyers make a second copy of the documents in their platform and leave behind all of the rich context present in the source systems. These approaches can be useful for task productivity, but they stop short of the kind of litigation intelligence, substantive case work requires. Our solution is different because it holds both directly. The case law itself and the full Ediscovery record metadata and work product included and reasons across them. The same is true of how litigation teams work day to day. Legal research, the document record, work product and the docket typically sit in separate systems today.

So any question that touches more than one of them becomes a manual exercise. Pull the law here, pull the documents there, stitch them together by hand and do it again each time. That's exactly why Eric described this as DISCO building a single pane of glass for the litigation team, bringing all of it into one place means that stitching happens once and the context built answering one question carries straight into the next instead of every person on the case team rebuilding the same connections over and over. The same logic is why persistent context matters so much. Litigation is not a single question-and-answer exchange. It's a matter that evolves over months or years through new productions, depositions and rulings.

DISCO carries that context forward automatically so its understanding of a matter compounds as the case develops. And this memory is what makes our solution so valuable because it drives better outcomes, not a faster answer to one question, but an answer that reflects everything the case has taught the system so far. Right now, we're in a pilot learning phase. We're running the pilot with a small hand selected group of customers on live matters because what matters to us right now is depth of engagement, understanding usage patterns, how capabilities resonate and what to prioritize on our road map.

Here's where we think it goes. Right out of the gate, our Unified Litigation Solution helps the case team answer important legal questions and develop case strategy, leveraging deep integration to both the facts and the law. This is backed by proprietary workflows that we've built specifically for litigators and we've filed 3 provisional patents that demonstrate the novel and innovative approach we've taken in building this solution. Based on our road map, we intend to iteratively add new capabilities such as the ability to maintain live case artifacts such as proof tables, witness materials, motion outlines and discovery plans that update automatically as the matter progresses. This supports the full case team working together, including outside counsel and in-house counsel in the same shared workspace. This will move quickly.

Our vision is ambitious, and the customers and partners we have engaged are excited about what we are building, and that is why we are treating this as one of the most important long-term investments we can make as a company. We'll share more on timing and availability as our plans evolve. This same learn first scale fast approach is exactly how we approached Advanced Research, which will roll out to all customers in the next few weeks. Advanced Research is an agentic AI capability built deeply into our Ediscovery products that goes well beyond simple question and answer. It performs multistep reasoning across a matter's full evidentiary record to help litigators develop a deep understanding of their evidence.

The customer response during testing has been powerful, as Eric noted, and we are excited to be ready to roll this out to customers. None of this happens without the work we've done for a decade on the Ediscovery platform itself. Just as important, our own engineering organization has been shifting to an AI-first way of building software over the past several quarters. And that shift in velocity is part of what has led us move as quickly as we have on the new Unified Litigation Solution, Advanced Research and Auto Review all at once. And underneath all of it, we keep investing in our ability to ingest and structure increasingly complex and modern data types because none of this works, including our new solution, if the data processing platform can't turn raw data into something a litigator or an AI system reasoning on their behalf can actually use.

Taken together, we believe this is the clearest evidence yet that our AI native stack is compounding. The solutions we've built, the customers who trust us with their largest and most complex matters and the data that flows through DISCO every day are what gives us confidence in the future. With that, I'll hand it over to Aaron.

Thank you, Richard. Q2 results reflected continued progress in the areas Eric and Richard just walked through. Total revenue was $43.1 million, up 13% year-over-year and software revenue was $36.8 million, up 13% year-over-year. Services revenue was $6.3 million, up 18% year-over-year, driven primarily by strength in professional services and Auto Review-related managed review work tied to our largest matters. I want to touch on some of what Eric mentioned in more detail. First, DISCO platform was our biggest driver of quarter-over-quarter software growth. Customer adoption is strong, matter count is growing rapidly, and our pricing is helping increase the lifetime value of each matter.

Second, Auto Review had a strong quarter. In Q1, we mentioned that some customers evaluating Auto Review were choosing the traditional route instead based on their AI readiness and comfort level. In Q2, that trend reversed. Auto Review set a new revenue record, driven by both record number of auto reviews executed and a growth in the average size of Auto Review to date. We are delighted by this result, but we are still in the very early stages of adoption. Turning to profitability metrics. As a reminder, unless otherwise specified, references to gross margin, operating expenses and net loss are on a non-GAAP basis, and adjusted EBITDA is also a non-GAAP financial measure.

Gross margin in Q2 was 76% compared to 76% in the prior year. Sales and marketing expense was $15.7 million or 36% of revenue compared to 36% of revenue in the prior year. Research and Development expense was $13.4 million or 31% of revenue compared to 31% in the prior year. The dollar increase reflects continued investment in our Unified Litigation Solution, Advanced Research and Auto Review to capitalize on the legal industry's AI transformation. General and administrative expense was $7.7 million or 18% of revenue compared to 19% in the prior year. Adjusted EBITDA was negative $3.4 million in Q2, representing an adjusted EBITDA margin of negative 8% compared to negative 7% in Q2 of the prior year. Net loss in Q2 was $3.6 million or 8% of revenue compared to a net loss of $2.8 million or 7% of revenue in Q2 of the prior year. Net loss per share was $0.06 compared to $0.04 in Q2 of the prior year.

Turning to the balance sheet and cash flow statement. We ended Q2 with $101.4 million in cash and short-term investments and no debt, maintaining our strong financial position. Operating cash flow in Q2 was negative $1.1 million compared to negative $4.2 million in Q2 of the prior year. I also want to reiterate a dynamic that I mentioned last quarter. If DISCO platform continues to perform above our expectations, we may see some short-term revenue impacts from lower ingest fees. We expect it will be more than offset over time by the DISCO platform's ongoing fees and by larger and longer and more complex matters that come with it. We believe this will be a positive trade for the long-term revenue profile of our business. We have not seen a meaningful drag on our results related to this new dynamic to date, but we may in the future.

Turning to guidance. For the third quarter of fiscal year 2026, we are providing total revenue guidance in the range of $43.75 million to $45.75 million and software revenue guidance in the range of $38.1 million to $39.1 million. We expect adjusted EBITDA to be in the range of negative $1.75 million to negative $0.25 million. For the fiscal year 2026, we are raising our total revenue guidance to a range of $172 million to $179 million and software revenue guidance to a range of $147.5 million to $152.5 million, reflecting an increased confidence in the second half of the year. We're updating full year adjusted EBITDA guidance to a range of negative $8 million to negative $5 million. We continue to expect to be adjusted EBITDA positive in Q4.

[Operator Instructions] Your first question comes from the line of Scott Berg with Needham.

I guess I got a couple. Let's start on Unified Litigation Solution that you all are seemingly quite excited about. I guess what is the customer adoption cycle of this platform look like? Is this an add-on to maybe an already existing Ediscovery customer, can you sell this net new? Do you have to have Ediscovery maybe implemented first since it's kind of leveraging that data? And then how do you think about pricing for that solution? Because off the top of my head, it sounds like maybe a consumption-based pricing model, which is the primary mechanism for your Ediscovery solution might not be the right pricing model for this, but would love to hear if that's maybe accurate or not accurate.

Yes. Let me get started on that question. I can tell you a little bit more about how we're thinking about the new solution commercially. As I mentioned in the prepared remarks, we are deliberately in a learning phase, right? And that's the way you should think about where we are today, right? We're running this pilot. We've got a small hand-selected group of customers. And they're deeply engaged on live matters, right? This isn't a proof of concept or something that people are just playing around with the demo. This is a real product. But we're intentionally right now not putting a firm pricing or monetization or timing plan in the market led because we want to learn, right?

We want to see how -- what we've developed and that we are excited about really plays out like how the capabilities resonate, how the customers' commercial needs and their business models should shape the way we should think about the best way to engage them commercially before we lock anything in. This is an ambitious multiyear bet and you should definitely hear today's announcement as the first step of a large vision, not the complete answer ready today. As we get further into the pilot and subsequent calls, we'll have much more to share on packaging and timing and certainly update you. But I wouldn't build revenue from the solution into your models yet.

Yes, certainly not for 2026. But Scott, I think the other thing I would just say is that this isn't something we're building because we think it's cool. This is something that our customers want. We're in a situation where our law firm customers, corporate clients are demanding that they get more value from their outside counsel spend, and they want to understand how the outside counsel can differentiate themselves. And DISCO is in a unique position to truly change the legal industry like no other AI company can. There's all this interest right now in legal AI, but nobody else in the market has the combination of the facts, the case law, the deep litigation experience and then the agentic AI capabilities that we've got here at DISCO.

Although other AI companies in the legal space are trying to help lawyers become more efficient. We're doing that, too. But in addition to that, we're helping litigators win. And we're incredibly excited about it. So we're being very thoughtful about the way we build out this capability with our customers and the way we're going to commercialize it. But we think this is an enormous opportunity for DISCO and it's really the evolution of the company moving forward.

Got it. Helpful. And then for my follow-up question on your fourth quarter guidance, or at least the implied fourth quarter guidance since we're now get to see the back half. It does assume that there's an acceleration in subscription revenues. You seem pretty confident around sales trends and some of the usage trends, especially upmarket with larger customers over the last quarter or 2. But I guess, where is the confidence coming into play about some reacceleration in those revenues in the fourth quarter?

Yes. I think simply, Scott, thanks for the question on it, too. But when you look at our guide, you are correct. We're raising the guide for the year and the confidence in raising that guide really comes from 3 things. The first thing is execution. Year-to-date, our execution against the strategy around more large matters from our largest customers. We're seeing the success in that. We are expecting that to continue. The second piece is DISCO platform. We talk about how we're excited with the trends we've seen, the adoption we've seen at DISCO platform on not just it's exceeded our expectations, but it's exceeded our expectations from a pricing of volume.

And so all the elements that we're expecting, we're seeing deposits come through there. And then on the third piece, is Auto Review. We talked about that as well. We're seeing positive traction there. It seems like AI readiness for our customers is improving, and we're also building new capabilities that make it easier for them do an Auto Review, and we're going to continue to make those investments as well. But I think those -- the combination of the 3 things is what's driving our confidence in the guide.

Your next question comes from the line of DJ Hynes with Canaccord Genuity.

This is Ryan on for DJ. So I guess I understand the difference between the transaction and legal work that you see with some of these legal AI start-ups and the more privacy concerns with Ediscovery solutions. But a trend we've been seeing across software is this general openness with these platforms. Do you foresee as they expand their platform, you expand yours and customers adopt these platforms more, you will have to open up your platform to them?

Let me try to answer that for you. I think it's important I focus on why it's important, particularly for our customers that we go deep, right, and that we have comprehensive access to the information that matters to them because litigation is just fundamentally more complex than all of the other transactional advisory legal work that is often the focus of general legal AI tools, right? It's adversarial, right? And as I said in the prepared remarks, right? The facts could be buried across millions of documents, not just in a small subset of some hot documents that might have been identified, right? And so you need to be able to have the depth of the facts. You also need to have that deep access to the legal corpus we have, right?

And the ability to build on it and train across it, not just access it one legal decision at a time, one statute at a time. And when you couple those 2 things, the real deep depth and the focus that we're putting on litigation and our ability and our proven leadership in developing AI over the last decade, that's what Eric was talking about. You have the formula that differentiates this 4 litigation teams, right? And litigation teams need something different, and it's going to be DISCO that delivers the thing that is most fit for their needs and why we're so excited about how impactful it will be for the litigation professionals that are going to use it.

Yes, let me just add into that too, real quick. Look, there's technology opportunities and limitations, but then there's also just business strategy opportunities and limitations. And the reality is the companies that own data aren't always extremely willing to give up that data. And one of the general AI companies in the legal space has a partnership with one of the big data providers for case law and the CEO for the case law company was interviewed last week and said that they've only got access to 1% of the data. I think that's what you have to think about. The reality here is we've got access to all of the data, and we've got control over how to use that data, which is a big, big differentiator for us.

Okay. Got you. And then a quick follow-up. So you guys have spoken before about how law firms are a channel into corporate legal department customers. So obviously, large matters are growing, but can you maybe just describe what that expansion motion into those corporate legal departments looks like compared to maybe your more traditional analog customers.

Sure. Yes. I mean, look, I think as you know, we think there's tremendous upside opportunity within our customer base. And if you think about our law firm customers, we talked about -- well, overall, our large customers, we've now got 354 customers that spend over $100,000 with us over the last 12 months. That's up 15% in terms of the revenue, it's up to 77% of our total revenue. It's significant. We continue to improve in that customer base and yet we've got these very large customers where we might still only have 15% to 20% of their wallet share.

And so what we've done is we've created an integrated go-to-market approach, where sales, marketing, sales development, customer success, really every piece of our go-to-market team is focused on helping enable our law firm customers to market DISCO internally to their case teams and then also ensures that those case teams can market both DISCO and their law firm services together to their corporate customers. And so honestly, we made great traction, but I think incredibly -- there's an incredible upside that we haven't even really seen flourish yet. We're making progress, but there's much, much more upside really through enabling these law firms to sell on our behalf. And the great news is, I was in London last week. I met with 7 different clients, and I had several of them talk about how they want DISCO to help them market to their customers because their corporate clients are asking how they're going to differentiate themselves.

How are they going to leverage AI to be more efficient and deliver better outcomes. And like I've never seen before, law firms can tend to be really conservative, particularly from a marketing standpoint, but that's changing. They want us to help them market them because they need to differentiate themselves and show value to their corporate clients.

Your next question comes from the line of Mark Schappel with Loop Capital Markets.

Eric, you called out the strong adoption of the DISCO platform in your prepared remarks. I was wondering if you could just talk a little bit more about how much of that traction you're seeing is coming from new logo wins versus, say, expansion with existing customers?

Yes. I mean, look, we're seeing -- it's a combination with DISCO platform. Again, the performance has been incredible. We hit our internal goals by June so we're obviously really excited. As you know, when we put out the new platform and pricing approach, there were multiple goals behind it. So one of them was to improve our consideration. And that's both for new customers, but also for new large matters within our existing customer base. There were many times in the past where customers wouldn't even consider us because they found -- our pricing seemed so much more expensive than our competitors. Even if it wasn't, it's just that our model was one that they couldn't understand very well. So we're seeing improved consideration. And then another goal was more -- better close rates because customers are electing to go with DISCO because they understand that our pricing is competitive.

And then the third thing is preserving our margins. There's been -- the theory was we were having to overdiscount because customers -- because they couldn't understand our pricing, we needed to discount more. And so we're very excited that all 3 of those goals so far for DISCO platform are coming through, and we're really excited about what we're seeing there. But it's actually for new customers and for existing customers.

Okay. Great. And you may have just answered my -- at least part of my next question here. But in the prepared remarks, a lot of discussion around products. That's understandable given the aggressive rollout over the past year. But on the go-to-market front, maybe you could just give us an update on what you're seeing on sales productivity wise and whether you're seeing, for example, new sales reps kind of reaching productivity faster as your AI capabilities are just easier to demonstrate.

Sure. Yes. Look, it's been a great several quarters here in terms of seeing sales efficiency improve. And some of it is the dynamic that I spoke about earlier, that we're doing a much better job with integrated go-to-market in terms of enabling our law firm customers to sell on our behalf, which allows us to get much more efficient. But another thing that allows us to get more efficient is large matters. We've spoken about this in the past. But what's great about selling larger and strategic matters, not only is it more revenue, but they stay on our platform a lot longer, which creates a lot of downstream opportunity for us. And we've had a lot better success with selling large matters because of a few different dynamics.

One is with you in every case. We've spoken about this before. But making sure that our customers for these very large matters, if they need services help, they understand that DISCO can help them from a services standpoint, whether it's ingesting data or managing a project that we're with them in every case, that has made a big difference. And another one is our ideal matter profile approach. So going after certain practice areas within law firms or certain industries within corporates, where we know DISCO delivers unique value and where the matters are bigger. So that's been a big boom for us.

And the other one is just innovation around our AI capabilities our agentic AI with Cecilia Advanced Research and Cecilia and Auto Review, all of those together, our GenAI capabilities have allowed us to win larger matters within our existing large customers. And then the last one, as Aaron noted earlier, was the DISCO pricing approach, the new DISCO platform, we rolled this out in January and the uptake has been fantastic, which has allowed us to be much more efficient. So it's -- all of this combined has allowed our salespeople to be much more productive.

There are no further questions at this time. I will now turn the call back to Eric Friedrichsen, CEO, for closing remarks.

Yes. Thank you very much. Look, I am extremely proud of the DISCO team for their excellent execution in Q2 and also for our customers for trusting us with their most important legal matters. I've said before that I think DISCO can be a 20% plus annual grower over time and Q2 reinforced that. We continue to make progress on the drivers that are going to help us get there, deepening our relationships with our largest customers, continuing to win the largest and most complex matters and driving broader adoption of our AI capabilities.

The solid execution that we've seen over the last several quarters has put DISCO in a position of strength and we believe that, that strength along with a tremendous set of assets, an established and growing AI platform, access to a comprehensive body of U.S. case law and a decade of experience creating innovative tools for litigators positions DISCO to define AI for litigation. I can't possibly be more excited and proud of our team for both executing on our core strategy and for innovating around AI for litigation. So really, really excited. I appreciate everyone's time today and the great questions, and we look forward to seeing you all next quarter.

This concludes today's call. Thank you for attending. You may now disconnect.

Ladies and gentlemen, thank you for standing by and welcome to CS Disco's First Quarter 2026 Conference Call. [Operator Instructions]

I would now like to hand the conference over to your first speaker today, Head of Investor Relations, Aleksey Lakchakov. Please go ahead.

Good morning and thank you for joining us on today's conference call to discuss the financial results for DISCO's first quarter of fiscal year 2026. With me on today's call are Eric Friedrichsen, DISCO's Chief Executive Officer; Aaron Barfoot, DISCO's Chief Financial Officer; and Richard Crum, DISCO's Chief Product Technology and Strategy Officer.

Today's call will include forward-looking statements within the meaning of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995 including, but not limited to, statements regarding our financial outlook and the future performance; our future capital expenditures; market opportunity, market position, product and go-to-market strategies and growth opportunities; and the benefits of our product offerings and developments in the legal technology industry. In addition to our prepared remarks, our earnings press release, SEC filings and a replay of today's call can be found on our Investor Relations website at ir.csdisco.com.

Forward-looking statements involve known and unknown risks and uncertainties that may cause our actual results, performance or achievements to be materially different from those expressed or implied by the forward-looking statements. Forward-looking statements represent our management's beliefs and assumptions only as of the date made. Information on factors that could affect the company's financial results is included in its filings with the SEC from time to time, including the section titled Risk Factors in the company's annual report on Form 10-K for the year ended December 31, 2025, filed with the SEC on February 25, 2026, and the company's quarterly report on Form 10-Q for the quarter ended March 31, 2026.

In addition, during today's call, we will discuss non-GAAP financial measures. These non-GAAP financial measures are in addition to and not a substitute for or superior to measures of financial performance prepared in accordance with GAAP. Reconciliations between GAAP and non-GAAP financial measures and a discussion of the limitations of using non-GAAP measures versus the closest GAAP equivalent is available in our earnings release.

And with that, I'd like to turn the call over to Eric.

Thank you, Aleksey. Good morning, everyone, and thank you for joining us. In the first quarter, DISCO delivered strong results across the board. With significant product momentum, continued traction with our largest customers in both software and services and strong underlying financial performance; we continue to drive accelerating and sustainable revenue growth. As AI permeates through the legal industry, law firms are looking for ways to boost their productivity, increase efficiency and deliver better outcomes in order to win more business as our corporate clients look to control litigation spend.

The key to success in this environment is trust. The trust is earned through security, enterprise scale and litigation specific capabilities necessary to help lawyers win on the largest and most complex matters. This is where DISCO shines. We are in an excellent place with unique capabilities to continue to be a disruptor and leader in AI for litigation and our progress in Q1 has helped further differentiate DISCO from both general purpose legal AI tools and those in the traditional Ediscovery space. In Q1, total revenue grew 14% year-over-year to $41.9 million and software revenue grew 12% year-over-year to $34.7 million.

This was the fourth consecutive quarter of accelerating growth in total revenue if you exclude the onetime contingent deal we recognized in Q3 of last year. We are very pleased to deliver strong software growth and to beat the high end of our total revenue guidance range. Adjusted EBITDA improved 32% to negative $3.5 million in Q1, also beating the high end of our guidance range. We saw strong Q1 performance in 4 key areas: first, increased wallet share among our biggest customers; second, growth of large multi-terabyte matters; third, continued adoption of our generative AI capabilities; and fourth, overall acceleration in the growth of data on our platform.

Regarding improvement with our biggest customers: in Q1 we increased the number of customers that generated more than $100,000 in total revenue during the last 12 months to 347. The revenue attributable to these customers during the last 12 months totaled $124 million representing 77% of total revenue over this period and 13% year-over-year growth. We are continuing to add multi-terabyte matters as more complex litigation comes on to our platform. In Q1 we saw an acceleration in net new large matters added, which is a very promising sign given that these matters generate more revenue, expand over time and last longer on our platform.

Continued adoption of our generative AI capabilities was driven by both Cecilia AI and Auto Review. DISCO is transforming high stakes litigation through an AI-native stack built on a decade of proprietary data innovation and purpose-built legal workflows. At its core, Cecilia's agentic intelligence allows legal teams to speak directly to their data, uncovering complex evidence in seconds rather than weeks. Cecilia Advanced Research is our new platform-native agentic AI capability, which is a breakthrough for Ediscovery and investigations.

It is capable of much more sophisticated autonomous reasoning that extracts deeper context, makes next level connections and delivers significantly more detailed and thorough results across even the largest data sets. We're currently in testing with select customers on live case data in preparation for a broader rollout to wait-listed customers next month. The feedback is fantastic. Customers instantly grasp how much more they can accomplish and see it as a real example of what other AI providers have only been promising. Increased adoption has also extended through to our AI-powered managed services, which deliver expert-level results at software scale economics.

The result is a secure enterprise-grade ecosystem that fundamentally redefines the speed and efficiency of modern discovery. DISCO Auto Review is a more accurate and leaner alternative to traditional review and an excellent example of our AI capabilities in action. As more law firms look for new revenue streams, Auto Review allows them to bring more of that work in-house rather than sending it to alternative legal service providers, moving review from a cost center to a profit center. This is a win-win-win for the client, for the law firm and for DISCO because it provides a clear ROI and better outcomes for the client while providing more differentiated revenue streams for the law firm and for DISCO.

Our Auto Review capabilities continue to lead the market in terms of speed and efficacy and we believe that as more and more firms consider AI for their review needs that we are very well positioned to capture that demand. Interest in Auto Review continues to grow. We are seeing more customers engage with us to evaluate how Auto Review can help with their larger matters and it has proven to be a strong driver for our Managed Review offering, which also enjoyed a strong Q1. That's a great example of how our AI capabilities combined with our customer value proposition of With You in Every Case are bringing more customers, more matters and more revenue to DISCO.

As far as overall acceleration of usage, we had a significantly better-than-expected launch of the DISCO platform in Q1. For context, the DISCO platform is our powerful industry-leading set of AI capabilities including Cecilia Q&A, auto timelines, document summaries, definitions and case builder bundled together in 1 solution with our Ediscovery capabilities on every matter. The DISCO platform gives customers everything they need to manage and win their matters for 1 competitive price. In the first 3 months, we've seen strong demand from customers with early adoption that has been much better than anticipated and we're equally pleased from a financial perspective.

While it's still in the early days, we're seeing some very encouraging trends from DISCO platform adoption, including larger matters, increased committed revenue, multiyear deals and growing AI adoption. These results demonstrate how much easier we've made it to do business with DISCO, something further proven by the strong customer demand. Continuing to grow the DISCO platform is a key driver behind expanding wallet share among our existing base of large customers with large matters. I always like to highlight a couple of real-world examples to illustrate the value that DISCO is delivering to customers.

These are examples of customers who have moved from important transactional relationships to strategic relationships that benefit us both. The first is Mound Cotton, a leading litigation boutique focused on insurance matters with a nearly 90-year history. Following the launch of our DISCO platform in the first quarter, Mound Cotton signed a 3-year enterprise agreement making DISCO the provider of choice for Ediscovery technology across their firm. The reasoning was simple. They wanted a strategic partner that combines secure cutting-edge AI technology with professional services and support that they need for their largest and most sensitive matters.

Mound Cotton conducted a broad review of potential partners in search of a comprehensive integrated solution before selecting DISCO and noted that it quickly became clear that DISCO was the better product for their clients and better experience for their attorneys. As a firm that closely works with large global financial institutions, Mound Cotton was drawn to DISCO's reputation for security, privacy and reliability. They said we have the luxury of being able to select the best-in-class solution and the unanimous verdict was that DISCO is a dramatically better product today and that the gap will only widen in the future.

As a firm with sophisticated clients that demand the best tools, DISCO is the right choice. We hear similar things from many of the top firms we work with. They need advanced secure technology paired with the expertise to help them get the most out of it to deliver results for their clients. The DISCO platform is making that easier than ever. A second example that demonstrates how we're building multiyear relationships because customers see our technology's potential is Reynolds Frizzell LLP, a generalist commercial litigation firm in Houston with a prominent energy litigation practice.

Reynolds Frizzell is one of our longest relationships. They've been using DISCO since 2015 and they also recently signed a multiyear enterprise agreement to expand their use of our technology across their firm. Reynolds Frizzell has taken a considered approach to new technology in the legal tech space thoughtfully vetting AI applications and focusing on technology specifically designed for legal use cases. As they looked into legal AI applications, DISCO was a natural place to start based on a decade-long relationship built on trust and collaboration.

Our Reynolds Frizzell partner said that they've used and evaluated a number of different AI legal tools and were especially impressed by DISCO Cecilia capabilities. We're excited to have it available for our cases, the partner said. This illustrates the power of our With You in Every Case value proposition. Our combination of advanced technology and expert professional services has made DISCO into an essential resource for Reynolds Frizzell and we're continuing to serve and grow this long-standing relationship into the future.

The stories about Mound Cotton and Reynolds Frizzell are just 2 examples of our strategy in action and there are dozens more every quarter demonstrating our ability to develop these relationships and dramatically expand them over time. All told, Q1 was a strong quarter for DISCO with continued growth in our core business, a better-than-expected launch of the DISCO platform and great progress with AI adoption. We believe this creates significant momentum for us throughout 2026 and beyond.

With that, I'll next turn it over to Richard to discuss how our recent product advancements and our product road map are shaping our longer-term view of the broader opportunity to provide powerful AI solutions for litigation. Richard?

Thank you, Eric. As Eric noted, we are incredibly excited about the customer response we're seeing to the DISCO platform and Cecilia Advanced Research. Both are important steps for us, but are really only the beginning of what we know is possible in litigation technology with our AI capabilities. The legal industry as a whole is in a period of significant change. Law firms face pressure to leverage new technology and consider changes to their business models. Corporate legal departments are expected to control spend while workloads are growing.

And everyone is grappling with the growing volume of new complex data leading to very real data management and fact-finding challenges. The general legal AI companies and even tools from foundational model providers are offering legal workflow solutions that address a wide range of transactional legal work such as redlining contracts, drafting memos and extracting information from a set of documents. Legal tasks for efficiency and automation to speed up work is truly valuable. Litigation is an entirely different game. It is significantly more complex.

Litigators are not asking for ways to work faster. They want solutions that shift the odds in their favor by delivering the crucial case intelligence that leads to victory. They need to win. And the expertise and precision required to deliver that requires scaled technology like we have developed at DISCO. Put another way, litigators need solutions that are purpose-built for the unique demands of the practice. It needs AI built for litigation. That's what we're building. At DISCO, we are continuing to extend our AI applications for our customers beyond traditional Ediscovery compliance to unlock both new strategic advantages and drive value across the litigation life cycle.

Let me explain that with some more detail. There are 2 high level outputs from Ediscovery, production compliance and a detailed understanding of the facts and evidence in context. Ediscovery has traditionally focused on production compliance because it is an important court mandated step in the litigation process with real consequences if you get it wrong. But it is also a necessary tick-the-box exercise with limited strategic value on its own. Most Ediscovery tools have made production more accurate, automated and efficient; but they have not made it more strategic or independently valuable.

This is where DISCO is different. Our ability to surface not just the facts, but the complete picture of context, intent and relationships between documents and data. DISCO goes beyond the required production to deliver a comprehensive set of facts and evidence organized, tagged, connected and understood in relation to the claims of the matter. The real value for litigators is in the mastery of those facts. The second piece is in the law itself and it's worth reminding everyone that DISCO holds a license for the full corporate of U.S. case law, statutes, regulations and court rules.

The facts of the case and the law of the jurisdiction together in 1 platform will be a powerful combination. We will share more about our vision and how it will come together soon, but let me turn back to what we're delivering right now. Three years ago DISCO first dramatically disrupted traditional Ediscovery with Cecilia AI. Our new Cecilia Advanced Research is an agentic AI toolset that leapfrogs the capability of those many copycat Q&A tools that have followed us by helping attorneys develop winning case theories during discovery and all along the litigation life cycle as the matter advances to settlement or trial.

Unlike simple Q&A tools, Cecilia Advanced Research functions as an intelligence agent performing multistep analysis across massive amounts of data to deliver defensible court-ready insights that litigators use to build winning case theory from day 1. Cecilia Advanced Research can generate work product for litigators as a case advances, build tighter and more compelling narratives with our integrated timelines functionality, streamline deposition and witness preparation and interrogate the record at trial and they do this all in 1 powerful integrated and secure platform from DISCO.

We're currently in testing with select customers on live case data in preparation for this broader rollout to priority customers on our wait list later this month and the feedback is fantastic. Customers instantly grasp how much more they can accomplish and see it as a very real example of what other AI providers have only been promising. The new era of AI, both generative and agentic, opens up a treasure trove of opportunity. You don't have to know how to use complicated technology. You just have to know how to articulate the outcome you're looking for, something that lawyers are already incredibly good at.

DISCO AI lets litigators focus on the output of the process, winning for their clients who hire them to deliver. We believe this means lower barriers to adoption, greater usage of our platform across the life of a matter and most importantly, better results for our customers and their clients. In the simplest of terms, at DISCO we are directly investing in our customers' competitive edge to help them win cases and grow their business. DISCO is the AI solution for litigators.

With that, I'll hand it over to Aaron.

Thank you, Richard. Q1 results were strong across all our revenue lines. We exceeded the top end of total revenue guidance in the quarter and came in above the midpoint of our guidance range in software. In Q1 2026, total revenue was $41.9 million, up 14% year-over-year while software revenue was $34.7 million, up 12% year-over-year. This was the fourth straight quarter of accelerating total revenue growth excluding the impact of onetime contingent software revenue recognized in Q3 of last year. Services revenue was $7.2 million, up 25% year-over-year.

To start, I want to touch on some of the dynamics we are seeing in our software business. As Eric mentioned, we saw strong traction in Q1 with DISCO platform. We are seeing more cases start on DISCO platform with more matters and gigabytes than we had expected through Q1 as customers see the obvious benefits of bundled products and all-in-one pricing. We expect these new larger matters will be a tailwind for our business in the coming quarters, but we could see variability as customers move from sets of individual products and ingest fees to the DISCO platform.

I also want to touch on our performance in services, which exceeded expectations in Q1 and was driven by growth of both professional services and our review business. We've discussed in the past the tremendous impact we believe Auto Review will have on the litigation workflow as more customers embrace AI adoption. While we have customers all along the spectrum of AI readiness, both new and existing customers are curious about Auto Review's capabilities. A further dynamic we are seeing is that as customers learn about both our traditional review and Auto Review, some choose to use traditional review as they consider broader AI implementation.

That dynamic helped fuel our strong services result in Q1. Turning to profitability metrics. In discussing the remainder of the income statement, please note that unless otherwise specified, all references to our gross margin, operating expenses and net loss are on a non-GAAP basis. Adjusted EBITDA is also a non-GAAP financial measure. Our gross margin in Q1 was 75%, consistent with 75% the prior year. As we mentioned before, our gross margins fluctuate from period to period based on the nature of our customers' usage, for example the amount and types of data ingested and managed on our platform.

Sales and marketing expense for Q1 was $14.8 million or 35% of revenue compared to 36% of revenue the prior year. The year-over-year dollar increase was driven by personnel costs as we invest in our go-to-market capabilities. Research and development expense for Q1 was $12.9 million or 31% of revenue compared to 33% of revenue the prior year. Research and development increased year-over-year primarily driven by higher personnel costs as our team continues to focus on AI and platform development. General and administrative expense in Q1 was $8.6 million or 21% of revenue compared to 23% of revenue in Q1 of the prior year.

General and administrative expense were relatively flat year-over-year. Adjusted EBITDA was negative $3.5 million in Q1 representing an adjusted EBITDA margin of negative 8% compared to an adjusted EBITDA margin of negative 14% in Q1 of the prior year, also a 600 basis point improvement. We are pleased with this progress and the fact that adjusted EBITDA exceeded the high end of our guidance. Net loss in Q1 was $4.2 million or negative 10% of revenue compared with a net loss of $4.9 million or 14% of revenue in Q1 of the prior year. Net loss per share for Q1 was $0.07 compared to $0.08 per share for Q1 of the prior year.

Turning to the balance sheet and cash flow statement. We ended Q1 with $103 million in cash and short-term investments and no debt, maintaining our strong financial position. Operating cash flow in Q1 was negative $11.7 million compared to negative $10.5 million in Q1 of the prior year. Turning to our guidance. For Q2 2026, we're providing total revenue guidance in the range of $41.5 million to $43.5 million and software revenue guidance in the range of $36.1 million to $37.1 million. We expect adjusted EBITDA to be in the range of negative $4.5 million to negative $2.5 million.

For fiscal year 2026, we are increasing our total revenue guide to the range of $169.25 million to $178.75 million and software revenue guidance to the range of $146 million to $152.5 million. We expect adjusted EBITDA to be in the range of negative $8 million to negative $4 million.

Now I'd like to turn the call over to the operator for Q&A. Operator?

[Operator Instructions] Your first question comes from the line of Scott Berg with Needham.

Probably a question for Eric or Richard, I wanted to start off with I think the A topic in the space in the quarter. There's been a lot of questions we feel it from investors on the ability for customers within the litigation space to use some of the tools that have been released on the frontier large language models out there and you all addressed that a little bit in your prescripted remarks. But I think the question in all of that is does it actually disrupt sales cycles or maybe how your customers are using the product during the quarter as they maybe "tried" or wanted to evaluate those technologies or do you really see it maybe more as a nonevent in your operational activities?

Scott, this is Richard. Thanks for that question and I think it's an important dynamic to unpack a little bit. It's certainly written about a lot that generative AI has the ability to commoditize some of the simple steps that lawyers do, right, whether it's summarizing documents or running a search or drafting. But I think it also in some ways elevates the intelligence layer where DISCO, right, that's where we've invested, that's where we shine because the context that comes from the power of how DISCO's platform powers litigation brings evidence and facts to life for our customers, right? It's so much more powerful than what any large language model or a tool that's simply built on top of that large language model could ever do even with the same data. And so I think what we're seeing is our customers realizing that Ediscovery actually presents a real shift from just a simple tool to perform a job into something that with DISCO can give them a strategic advantage.

Yes. Scott, I'll add on to that. This is Eric. No, we haven't seen any slowdown at all in sales cycles related to these new tools that have come out. In fact if anything, it's helped us drive AI adoption because lawyers, overall law firms are much more interested to see how AI can impact them. And we've got AI that can drive incredible ROI and help provide better outcomes ultimately for customers' clients. So the short answer is no, we haven't seen any negatives at all. It's been very positive for us.

Excellent. And then from a follow-up perspective, you all commented that the DISCO platform I think saw better interest and adoption in Q1 than maybe what you had initially anticipated there. But in your conversations with customers so far, I guess what have you found in terms of pricing and use relative to, I don't know, a customer that was just using Ediscovery before. If you can help us understand maybe what that opportunity or journey is like to take a customer from what's historically been a single solution on the DISCO platform to the all-in opportunity there. I think that would be really helpful.

Really demand for AI is what drove far better-than-expected results with the DISCO platform adoption. Certainly, there was some pent-up demand from both our customers and our sales teams. They were excited for the opportunity to have Cecilia AI, Case Builder and all of our core Ediscovery capabilities integrated across all of their matters. So that's the main driver. But also as you remember, our old pricing model was hard to understand and it made some customers feel that we were much more expensive than the competition especially for larger matters when we were actually pretty similar.

And so a shift to more of an apples-to-apples pricing model has really increased our consideration for new matters and for new customers and it's made a big impact already. I mean we've seen some very big and complex matters starting the DISCO platform right here out of the bat in Q1. We've seen increased revenue commitments. We've seen longer-term agreements from some of these customers. So DISCO platform is off to a great start and I'm really optimistic about the future.

Our next question comes from David Hynes with Canaccord.

Eric, I wanted to ask a big picture kind of industry implications question related to the AI-driven advancements we've seen in legal tech. Do you think it makes it so that the largest firms are able to take on more so that they own kind of more of the space or does it level the playing field so that smaller firms are able to be more competitive? And I guess what are the implications of all this perspective change for DISCO?

Yes, I'll get started and others can feel free to chime in. But look, I think this is an opportunity for law firms to generate more revenue. The whole legal industry right now is rethinking their business models. And what AI can do is give the opportunity for these law firms to be able to take more business in-house that they were previously sending out to alternative legal service providers. If you think about particularly when it comes to the review process, this low level, low dollar work that was fairly mundane tasks that law firms didn't really feel like fit their model for the most part in how they wanted to provide value to their customers.

They were sending it out to these alternative legal service providers doing that human work and part of that was they were missing out on the revenue. The other part of it was they were losing the context of all that great work. And so now the fact that they have the opportunity to leverage much, much better technology with generative AI and products like Auto Review to be able to bring that business back in-house, add extreme value on top of it in terms of legal judgment and generative AI consulting services and keep the context in-house to really help their lawyers go drive case strategy. It's a real game changer.

Ultimately, the law firm can generate more revenue. That's good for DISCO. But also the end client can save money and get better outcomes. So ultimately, I think this is a big shift and a big opportunity for each of us.

Yes. Makes sense. Aaron, a follow-up for you. So the $100,000-plus net add number was particularly strong this quarter, but software revenue has held more or less flat the last few quarters. Can you just help me understand that dynamic? Are the customer adds a leading indicator and software revenue should follow or is the uptick in folks moving over that spend threshold just services driven? Like how should we think about this?

I think when you look at the quarter and you look at the movement of the $100,000 customers, we're certainly happy that sequentially it grew 5% quarter-over-quarter and I would definitely characterize that as a leading indicator. And the reason for that is obviously the matter comes in, it ingests and then it expands and moves on to the platform. So that's the dynamic that happens and that's why it's a leading indicator. I'd also add though with that when you think about it as a leading indicator, there's going to be -- there's movements both ways, right? You have matters coming in, you have matters going off.

And so you're always in a usage model looking at the triangulation of both. And so it is obviously having a strong quarter usually is a good leading indicator of those matters coming in, but it also depends on what's coming out and that's what goes into our models. I think when you look at the quarter, relatively speaking, one of the other elements you asked about kind of with the growth is, and I touched on this in my prepared remarks, with Auto Review. We're super happy with the traction we've seen. It's actually brought us new customers' matters. It's brought us matters from existing customers. And in bringing those in, what happens is those customers come in at varying states of AI readiness.

And so as it comes in through the pipeline, we've seen a handful of those go and convert and become Managed Reviews. And so that helps drive -- so Auto Review actually helps drive part of the fee on the services line for the quarter, which is why we're very proud to come at the upper end of the range there. At the same time, we're happy that some of those came in, became Managed Reviews; but subsequently, those customers have come back to us with other matters and chosen Auto Review. So I think what you're watching, and Eric kind of alluded to this too, is law firms are becoming more and more comfortable with AI and how it plays and so we're watching that closely.

Yes. I think it also just speaks to in every case, customer value proposition, right? So we're with our customers in every case. And if they choose at a certain point because we're not quite ready to use Auto Review, we can leverage our AI managed services and our Managed Review to really help them in the short term. So overall, I'm incredibly pleased about the revenue, 14% revenue growth for the business this quarter.

Yes, yes. And just so we're all perfectly clear. Auto Review falls into the software line and Managed services obviously falls in the services line. Is that correct?

Correct. And Auto Review actually has 2 components today. It actually has -- part of it sits in software and part of it sits in services. And the reason for that is that actually goes to some of the familiarization part of it as well. What happens is today, a customer, we might actually help engineer the prompts for them. So that's still manual work that we do as we set them up. But I think as customers become more familiar over time, that prompting will not be required and so it will become fully -- it will be truly software revenue. But today, it sits partially in software and partially in services.

There are no further questions at this time. I will now turn the call back to Eric Friedrichsen, CEO, for closing remarks.

Yes. Thanks, everyone. Thanks for the questions. Q1 was a strong quarter for DISCO. The demand for our AI capabilities drove better-than-expected results for the launch of the DISCO platform. We're already seeing benefits from our new pricing model, which was designed to increase consideration, to improve win rates, to reduce discounts, to improve stickiness and ultimately to provide more value to our customers. I've said before and I'll say it again that I believe that DISCO can be a 20%-plus grower over time. In the first quarter, we made continued progress on the key drivers that are going to make that happen; things like increasing our share of wallet with our large customers, acquiring larger matters and accelerating AI adoption.

So all 3 of those demonstrate that our strategy is working. And along with our product road map and where we're going next, DISCO is really poised to be the leader in AI for litigators. We've increasingly transformed high stake litigation through our purpose-built legal workflows. We are very much focused on litigation not general legal. And when you combine our strategy with our path to reach adjusted EBITDA profitability in Q4 of this year, we believe DISCO is on an excellent trajectory today and we're positioned for the future as our solutions become increasingly the standard to help litigators win. So thanks for your time today. We'll see you next quarter.

This concludes today's call. Thank you for attending. You may now disconnect.

Ladies and gentlemen, thank you for standing by, and welcome to CS Disco's Fourth Quarter and Fiscal Year 2025 Conference Call. [Operator Instructions]

I would now like to hand the conference call over to your first speaker today, Head of Investor Relations, Aleksey Lakchakov, please go ahead.

Good morning, and thank you for joining us on today's conference call to discuss the financial results for Disco's fourth quarter and fiscal year 2025. With me on today's call are Eric Friedrichsen, DISCO's Chief Executive Officer; Aaron Barfoot, DISCO's Chief Financial Officer; and Richard Crum, DISCO's Chief Product Technology and Strategy Officer.

Today's call will include forward-looking statements within the meaning of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995, including, but not limited to, statements regarding our financial outlook and future performance, our future capital expenditures, market opportunity, market position, product and go-to-market strategies and growth opportunities and the benefits of our product offerings and developments in the legal technology industry.

In addition to our prepared remarks, our earnings press release, SEC filings and a replay of today's call can be found on our Investor Relations website at ir.csdisco.com.

Forward-looking statements involve known and unknown risks and uncertainties that may cause our actual results, performance or achievements to be materially different from those expressed or implied by the forward-looking statements. Forward-looking statements represent our management's beliefs and assumptions only as of the date made.

Information on factors that could affect the company's financial results is included in its filings with the SEC from time to time, including the section titled Risk Factors in the company's quarterly report on Form 10-Q for the quarter ended September 30, 2025, filed with the SEC on November 5, 2025, and the company's upcoming annual report on Form 10-K for the year ended December 31, 2025.

In addition, during today's call, we will discuss non-GAAP financial measures. These non-GAAP financial measures are in addition to and not a substitute for or superior to measures of financial performance prepared in accordance with GAAP. Reconciliations between GAAP and non-GAAP financial measures and a discussion of the limitations of using non-GAAP measures versus their closest GAAP equivalent is available on our earnings release.

And with that, I'd like to turn the call over to Eric.

Thank you, Aleksey, and good morning, everyone. Thank you for joining us. Right here at the start, I want to welcome aboard Aaron Barfoot as DISCO's Chief Financial Officer. Many of you saw our announcement in December, and we're thrilled to have Aaron on board. His deep experience in enterprise software and AI-driven business transformation at several industry-leading companies, along with his inspiring leadership approach is a perfect match for DISCO. Aaron has hit the ground running and has already become a great partner on DISCO's journey to revolutionize the Ediscovery industry.

Speaking of that journey, I can tell you that I have never been more confident of DISCO's role as the disruptor in this industry and our ability to help our customers drive better outcomes for their clients and their litigation matters.

DISCO was built from the ground up on cloud-based AI native technology specifically designed for the rigors of high stakes, complex litigation. Unlike general purpose AI tools, DISCO is built by lawyers for lawyers across massive volumes of complex and sensitive data with privilege controls, audit trails and litigation-specific workflows that lawyers can stand behind in court.

In order to best understand this, you really need to draw a mental picture of the 4 layers of our AI native stack. At the foundation sits DISCO's proprietary data layer, which is the result of a decade of innovation on data, machine learning and artificial intelligence with inference engines that power the industry's fastest and most advanced Ediscovery platform.

Built on top of that foundation is DISCO's core Ediscovery solution with its integrated workflows that are purpose-built for litigation professionals and trusted across the most complex high-stakes matters in the world.

Our third layer brings in generative AI with Cecilia. It answers complex questions in natural language and surfaces key evidence in seconds, connecting complex and nuanced concepts across different types of data to dramatically accelerate evidence finding and document review. This is the layer where we recently announced agentic capabilities such as advanced research.

Cecilia allows lawyers to speak to the data like never before possible. At the top of the stack, DISCO's managed services layer is powered by auto review, bringing generative AI to the work traditionally carried out by large human review teams and thus delivering managed service expertise at software scale and economics. The result is a coherent AI native stack underpinned by the enterprise-grade security, compliance and auditability that litigators require.

Our opportunity to disrupt the industry was a key driver in the strategy we built and began executing after I joined DISCO in 2024. Coming into 2025, DISCO set high goals for progress against our new strategy, and I am very proud of the team's results.

In Q4, total revenue grew 11% year-over-year to $41.2 million and software revenue grew 14% year-over-year to $35.1 million. This was the third consecutive quarter of accelerating growth of both total and software revenue, excluding the onetime contingent deal we recognized and called out specifically in the previous quarter. We are very pleased to beat the high end of the guidance range that we provided for both software and total revenue.

Adjusted EBITDA was negative $2.2 million in Q4, representing an adjusted EBITDA margin of negative 5% compared to an adjusted EBITDA margin of negative 12% in Q4 of the prior year. Full year 2025 total revenue was $156.8 million, up 8% year-over-year, while software revenue was $134 million, up 12% year-over-year. Full year 2025 adjusted EBITDA was negative $10.2 million, a margin of negative 7% compared to a margin of negative 13% in 2024.

While I continue to be proud of the end results, I'm even prouder of how we got there as it continually reaffirms that our strategy is working as we drive to durable growth and sustainable profitability over time. The main contributors to our performance included overall growth in usage on our platform, increases in large matters, growth with large customers and acceleration of adoption of our generative AI capabilities.

In Q4, we set record highs in total terabytes on our platform with accelerated year-over-year growth. We finished the year with double-digit growth in multi-terabyte matters and with revenue growing over 30% from those matters year-over-year in Q4. We increased customers that generated more than $100,000 in total revenue during the last 12 months to 330. The revenue attributable to these customers totaled $119 million in 2025, representing 76% of total revenue.

Additionally, we saw significant acceleration in the adoption of our generative AI capabilities, including Cecilia AI and Auto Review. This resulted in Q4 year-over-year growth of over 600% attributable to these features. And it's important to point out that when customers choose to use our Gen AI capabilities, they're also choosing to use our integrated AI native Ediscovery offering. So the impact is even greater.

One year ago, we introduced our new customer value proposition with you in every case. This approach prioritizes successful outcomes for our customers across the most important cases by leveraging the power of our platform, AI capabilities and expert services teams. We are bringing AI to lawyers and to the Ediscovery industry that has for decades, relied on inefficient human-powered document review, which is often outsourced to armies of contract attorneys and paralegals.

Our core DISCO platform with Cecilia AI, including our newly announced agentic capabilities and Auto Review, put the power of industry-leading technology back in the hands of the clients and their law firm partners. Think of agentic Cecilia as a senior investigator that lives inside your data. It reasons, it performs multistep operations. It links nuanced concepts between different documents and topics to paint a story complete with sources, not just give simple answers. It moves you from the what of the case to the why, connecting dots between disparate information points to identify patterns and facts that a human might miss, and it does this across even the largest data sets.

Auto Review, our Gen AI document review, brings that concept to the next level by taking that legal intelligence and applying it across incredibly large data sets. And we've used millions of documents with precision and recall that consistently outperforms human teams in a fraction of the time.

I think it's always best to bring things into perspective when you hear how our customers are specifically benefiting from such a partnership with DISCO. One example is of a long-term large law firm customer who recently chose DISCO for a highly complex case for a large construction conglomerate under an urgent time line. They chose DISCO's Ediscovery, Cecilia AI and Auto Review. The data set is multiple terabytes and after using Cecilia AI in concert with our other capabilities, narrowed the review population to about 550,000 documents.

After the initial prompt setup, our AI-powered Auto Review completed the review in just 2 days, delivering 98% precision and 97% recall, results that are not only superior to industry accepted human review standards, but provide the statistical defensibility our clients require for court-mandated productions.

To give a sense of the size of impact, in order to hit the same deadline with human reviewers, this would have taken a team of 70 people 4 weeks to complete. Instead, the client had the information they needed in record time with exceptional quality. This is a powerful example of how our Gen AI capabilities augment our core litigation platform and are transforming outcomes for our customers and why they decide to choose DISCO time and time again for their most important matters.

The second example that I want to mention is from one of the preeminent law firms, Osborne Clarke. While they were already a fantastic customer when I joined DISCO 22 months ago, they were using us primarily for smaller cases. They were very pleased with the DISCO platform, but we were unaware of the service that we provide to help on their larger and more complex cases. Over the course of this past year, DISCO has continued to build on the partnership with Osborne Clarke by providing strong client service, exceptional results and value for the money.

We made sure that our enterprise-grade software, leading AI technology and comprehensive services came through clearly in every interaction. So over the course of 2025, they more than doubled their matters with us, began using Cecilia AI, tapped into Auto Review for the first time and expanded their total spend with us by over 4x their 2024 total. Osborne Clarke is an innovation-focused law firm, and they are leveraging their partnership with DISCO to help enable them to deliver better outcomes for their clients.

The story of 2025 was one of executing on our strategy, investing in systems, processes and teams critical for our future, innovating on our core platform, Cecilia AI and Auto Review and accelerating growth and improving profitability. We started to see the benefits of our new strategy in action. That said, there's much more upside to come and to assure that our focus will be very much the same in 2026.

With that, I turn it over to Richard Crum, our Chief Product, Technology and Strategy Officer, for some exciting updates.

Thanks, Eric. In our daily conversations with customers, including at our recent Customer Advisory Board meeting, it's clear that AI is a powerful catalyst for the legal industry. The real excitement, however, lies in how we apply it. While general purpose AI tools offer interesting solutions for transactional work, our customers highlighted that they are not sufficient for high-stakes litigation.

Let me outline a few key reasons why purpose-built technology for AI for managing complex litigation is so important to our customers. Accountability remains with lawyers, not AI. Our platform is designed to leverage AI in a secure, compliant and defensible way. We provide a robust platform built on proprietary data architecture and strict privilege controls that manages data accessibility and establishes audit trails that law firms require to avoid malpractice risk.

When you are dealing with the most sensitive data, corporate trade secrets, internal executive communications, financial records, there is no room for error. Every insight generated by Cecilia AI is anchored by a source. Our Auto Review results exceed long-standing industry precision and recall benchmarks established for defensibility of a technology-assisted review in a court of law. Our customers know that DISCO software will provide them with the high-quality output they can trust their careers with.

Next, we are solving for scale. Many of the recently announced AI point solutions are great at helping lawyers with discrete legal work, such as drafting a memo or reviewing an agreement. This is useful to a legal professional, but it is not the same as supporting the legally mandated document review processes we support.

DISCO is used to manage multistep litigation-specific workflows across millions of documents. Litigators are connecting the dots between years of private data, including e-mail, Slack messages, PDF, audio, video and financial records to find the evidence and build their case. This is not a simple search problem. It is a massive multi-format data engineering problem that requires an industrial-grade backbone to ingest, secure and act on terabytes of data.

Further, litigation is a complex team-oriented workflow. Document Review is a process involving many stakeholders, law firm partners, corporate attorneys, associates, paralegal and service providers are all participants in a litigation. We provide the collaborative infrastructure that stand-alone AI tools are not built to handle.

And finally, our AI just works. With our newly announced agentic reasoning Cecilia Q&A capabilities, which we will be rolling out over the coming quarters, we have moved into the next level of legal insights. Cecilia is now an AI assistant that can find the deep connections across all types of evidence to answer why something is happening. The deep research mode for Cecilia can understand the question, create a plan, execute a multistage research flow, verify sources and deliver cited answers. Again, it works across millions of documents.

We are not offering AI wrappers that summarize text and [ flag ] risk. We have built agents that are deeply embedded in our litigation platform and help lawyers accelerate multistep litigation workflows. We are bringing our customers the capabilities they have been asking for like analyzing evidence for inconsistencies, building out detailed issue-specific time lines or figuring out the all-important who knew, what when questions.

We continue to push the boundaries of what our platform can do, both with the core technology that our lawyers require and the AI that makes them better at their craft. It is why our customers trust DISCO to lead the way in bringing innovation to legal tech.

Now I want to transition to a new topic. While we have been making so much progress in innovating our products, we have also been evolving our pricing model to unlock more value for customers. For example, we see positive reactions and repeat usage from customers once they try Cecilia AI, and we want to make sure we reduce barriers to Cecilia adoption as much as possible. Therefore, we are excited to announce that going forward, we are combining all of our powerful DISCO Ediscovery and Cecilia AI capabilities into a single offering, along with updates to our pricing and contracting approach.

Let me dive a little bit deeper into these 3 exciting changes that we are making to how we bring our products to market. First, and I think most excitingly, with our new pricing model, all of Cecilia AI will be included on every matter. This means that incredibly powerful and industry-leading tools like Cecilia Q&A, auto time lines, document summary, definitions and all the new skills we'll be adding to the Cecilia AI solution will be included on every matter under this new sales model. In addition, all of the capabilities of our case builder products will also be included. This brings our witness prep, deposition management and case story building tools together with our Ediscovery and Cecilia AI capabilities in one solution.

The DISCO platform gives customers everything they need to manage and win the matters for one competitive price. We are also evolving the way we price the DISCO platform. In addition to now offering all of the great technology I just spoke about for one simple per gigabyte rate, we are also adopting the industry standard approach to pricing. The formula will be based on the size of the customer data as it grows over time rather than basing our pricing on the initial data load size, which has often required us to discount our rates to meet competitor pricing levels, reducing both revenue and margin from full potential.

The unique approach we have historically used offered clients a level of cost certainty, but often also led to some customers wrongly viewing DISCO as more expensive compared to other similar solutions and caused us to lose out on sales opportunities.

Finally, we are updating our contracting options to better match the way our customers want to buy DISCO. We know there are good reasons why law firms and corporations desire to buy either matter by matter or by selecting a solution to standardize on. We also recognize that often the decision-making process can be influenced by different drivers. To meet the market, we will now offer simple contracting alternatives that make it easier than ever to select and do business with DISCO.

We have been testing this approach with select customers for the last 6 months and have received great feedback. Starting today, it is now generally available to everyone, and we are looking forward to discussing this new platform and pricing model with customers at Legalweek in New York in a few weeks. All these changes in new product solutions are driven by a deep understanding of how our customers' businesses are evolving along with the powerful technology we offer.

The long-term value derived from these changes will come in 3 ways. One, this new model will provide DISCO customers with all the tools they need to do their jobs better than ever before and elevate their legal craft and capabilities. We believe that when more customers see the full power and potential of the DISCO platform, there will be a natural acceleration in usage of both our core capabilities and AI.

Second, by meeting the market with a more simplified and industry standard pricing, we expect to see an increase in win rates with less discounting pressure. At full implementation, we expect that this will provide a revenue and gross margin lift for DISCO over the long term.

Third, by presenting customers with smart buying options, we make it easier to buy from DISCO and reward our most loyal customers who make spend commitments with our best rates on technology and services. Over time, this will grow our percentage of committed revenue, making our long-term revenue performance more predictable.

We are excited to bring this change to market and are confident it will reinforce DISCO as an industry leader, driving the future of legal technology by giving customers everything they need for a competitive price in a smart commercial model. These pricing model changes, coupled with our leap forward in product capabilities are part of our core strategy to grow wallet share with our largest customers and attract the largest and most strategic matters to our platform.

And with that, let me turn things over to Aaron.

Thank you, Richard. First and foremost, I'm very excited to be here with the team at DISCO as we revolutionize Ediscovery, litigation and accelerate DISCO's momentum. In this first month, I've been diving deep into the company, operations and vision. So far, everything I've seen reaffirms my thesis for joining DISCO, which is that the company has industry-leading technology and has the capacity to transform a historically services-oriented space into an AI-enabled software workflow. Capabilities such as Cecilia and Auto Review are central to that thesis.

Eric has built a strong leadership team that balances domain-specific expertise with technical know-how. I believe that with the right operational execution and financial discipline, DISCO has the potential to accelerate growth, produce robust free cash flow and generate attractive returns to our shareholders.

With that, I would like to discuss our results. In Q4 of 2025, total revenue was $41.2 million, up 11% year-over-year. Software revenue was $35.1 million, up 14% year-over-year. This was the third consecutive quarter of accelerating revenue growth, excluding the impact of onetime contingent software revenue recognized in Q3. Services revenue was $6 million, down 3% year-over-year, driven by a reduction in traditional review.

Full year 2025 total revenue was $157 million, up 8% year-over-year. Software revenue was $134 million, up 12% year-over-year. Services revenue was $22.8 million, down 8% year-over-year. The decline was attributed to a decline in our traditional review business. However, we are excited as Auto Review had strong growth in the first year sales and partially offset the decline.

We're pleased to see Auto Review show nice adoption this year. What's even more positive is we're seeing repeat usage. The Auto Review process involves the customer and our AI team developing a review prompt for Auto Review to execute across millions of documents. This motion results in services revenue in addition to the software revenue. As customers begin to move to the prompt process without DISCO support, more of this revenue will be purely software. Our traditional review product is still all in services.

We exceeded the top end of the guidance provided for the quarter across both software and total revenue. Looking back to the initial full year guidance we provided in February of 2025, we beat the high end of software revenue and came in near the high end of total revenue in that initial '25 guide. We accelerated the growth of our software business for the third year in a row from 3% in 2023 to 7% in 2024 and now 12% in 2025. We are exiting the year with momentum in usage growth and AI adoption. I firmly believe there's a lot more opportunity ahead.

I want to touch more on the usage dynamics that drove our growth in 2025. We saw accelerating growth in the transactional gigabytes and revenue on our platform, especially the gigabytes of complex multi-terabyte matters. These require an enterprise caliber approach to the sale, providing value to the customer through software, navigating complex objections and leveraging our services team to ensure success for the customer.

These factors combined to increase our software dollar-based net retention to over 103%. Total dollar-based net retention finished the year at 98%. We finished the year with 20 customers contributing more than $1 million in revenue, while our multiproduct attach rate was 19% at year-end, including our AI capabilities, leaving a large opportunity to expand within our existing customer base.

In discussing the remainder of the income statement, please note that unless otherwise specified, our references to gross margin, operating expenses and net loss are on a non-GAAP basis. Adjusted EBITDA is also a non-GAAP financial measure. Our gross margin in Q4 was 77%. Gross margin for the fiscal year 2025 was 76% compared to 75% in fiscal year 2024. As we mentioned before, our gross margins fluctuate from period to period based on the nature of our customers' usage. For example, the amount and types of data ingested and managed on our platform.

Sales and marketing expense for Q4 was $13.9 million or 34% of revenue compared to 37% of revenue in Q4 of the prior year. For fiscal year 2025, sales and marketing expense was $54.4 million or 35% of revenue compared to 39% of revenue for fiscal year 2024, a decrease of over $2.3 million year-on-year. The decline was primarily driven by a decrease in personnel costs and a reduction in marketing spend.

Research and development expense for Q4 was $13.0 million or 31% of revenue compared to 32% of revenue in Q4 of the prior year. For fiscal year 2025, research and development expenses were $48.4 million or 31% of revenue compared to 30% of revenue in fiscal year 2024, an increase of over $4.5 million year-on-year. This increase was primarily driven by an increase in research and development personnel spend as we continue to invest and innovate in our product capabilities.

General and administrative expenses in Q4 were $7.9 million or 19% of revenue compared to 20% of revenue in Q4 of the prior year. For fiscal year 2025, general and administrative expenses were $31.3 million or 20% of revenue compared to 22% of revenue in fiscal year 2024.

Adjusted EBITDA was negative $2.2 million in Q4, representing an adjusted EBITDA margin of negative 5% compared to an adjusted EBITDA margin of negative 12% in Q4 of the prior year. Adjusted EBITDA in fiscal year 2025 was negative $10.2 million, a margin of negative 7% compared to a margin of negative 13% in 2024.

Net loss in Q4 was $2.5 million or negative 6% of revenue compared to a net loss of $4.3 million or negative 12% of revenue in Q4 of the prior year. Net loss in fiscal year 2025 was $10.7 million or negative 7% of revenue compared to a net loss of $17.2 million or negative 12% of revenue in 2024. Net loss per share for fiscal year 2025 was $0.17 per share compared to $0.29 per share for fiscal year 2024.

Turning to the balance sheet and cash flow statement. We ended Q4 with $114.6 million in cash, cash equivalents and short-term investments and no debt. Operating cash flow in fiscal year 2025 was negative $14.9 million compared to negative $8.7 million in fiscal year 2024.

Now turning to the outlook. For Q1 2026, we are providing total revenue guidance in the range of $39.0 million to $41.5 million and software revenue guidance in the range of $33.75 million to $35.25 million. We expect adjusted EBITDA to be in the range of negative $6 million to negative $4 million. The decrease in Q1 2026 adjusted EBITDA relative to Q4 is primarily driven by increased employee costs, onetime expenses related to sales kickoffs, marketing campaigns and professional services. We believe we will be on track to achieve adjusted EBITDA breakeven by Q4 of 2026 as our revenue grows and as onetime costs in the first half do not reoccur.

For fiscal year 2026, we anticipate total revenue guidance in the range of $167 million to $177 million, and software revenue guidance in the range of $145.5 million to $152.5 million. We expect adjusted EBITDA to be in the range of negative $8.5 million to negative $4.5 million. The story of the coming year will be continued growth acceleration, driving us to adjusted EBITDA breakeven by Q4 of 2026.

Now I'd like to turn the call over to the operator to open up the line for Q&A. Operator?

[Operator Instructions] Your first question comes from Scott Berg with Needham & Company.

Two questions for me. Eric, I wanted to start off with the pricing and packaging changes. I guess, why now? You've been there, obviously, 22 months, why not maybe a year ago with what you've seen? And then how does that impact maybe existing customers and their current contracts? And I guess do you expect it to, I assume, positively impact deal cycles going forward, but any thoughts on deals that are in process? Is there any opportunity to disrupt or maybe accelerate those deals?

For sure, Scott. Thanks for the props on the quarter, too. It was a great quarter, and I'm super proud of the team. I'm going to let Richard here expound in a second. But in terms of the packaging and our pricing approach, look, we just saw an opportunity driven really by the demand from our customers. And so I'll let Richard talk a little bit more about how we work through that process.

Yes. Thanks, Eric. And you're right, the impetus for the pricing model changes that we talked about and that we're bringing to market start with listening to our customers who tell us, right, they want to use DISCO more and they want to use it on larger matters, but that they faced or faced some friction in selling it into some of the partner teams or in corporate teams because of the uniqueness of the way in which we had previously priced. And so we took that feedback and you couple with the vision that we have for DISCO, and that's what landed us on this new approach.

As I said in my prepared remarks, we've been out testing this, right? So we've been running this through with customers and signing deals based on this new model. And the feedback has been great. They're really excited about the inclusion of all of our tools and all of our AI into the core offering. And we're excited that it's going to reduce the friction to getting our customers access to that great technology, right? Making it easier to buy means they're going to have the best tools on more matters.

And it's great for DISCO, right? We do expect it will improve our win rates, help us win those larger matters that we've been growing with many of our customers, which ultimately leads to an improved sales efficiency and a higher lifetime value of matters because as we've talked about in previous calls, those larger matters last on the platform a lot longer. So we're real optimistic about the impact this new model is going to have on DISCO's performance.

Understood. And then from a follow-up question, Eric, you mentioned down there 22 months now. The company has accelerated its revenue growth rate for 2 straight years. Very, I'd say, very positive on -- especially on the software revenue line item. But as you've seen the business evolve with what you're looking at, whether it's product changes or the pricing packaging changes, how do you think about the intermediate-term growth rate of the company now? What does that look like to you? Is it at a rate higher than where you are today? Is it lower? Just help us understand maybe some of the industry dynamics and how you triangulate to what's the right kind of stable growth rate as you proceed forward.

Yes. I appreciate it, Scott. Look, I have been really proud of the team and the fact that we went from 3% growth to 7% growth to 12% software growth over the last 3 years. And I've said before that I believe the DISCO can be a 20%-plus grower, and I'm actually more optimistic than ever. I actually think we have 20% in our sights, not calling out a specific quarter or time frame that we're going to hit that, but we clearly have 20% on our sights, and I believe we can grow much faster than that.

So just getting to 20% plus growth, if you just look at the strategy that we're driving towards with our larger customers with larger matters and with more adoption of our generative AI capabilities, those things alone can help us get to 20% plus growth. As I've mentioned in the past, if you look at many of our largest customers, they are spending more than $100,000 with us and in some cases, more than $1 million with us. In many of those cases, we might only have 15% or 20% of their wallet share. So doubling down in that particular strategy and driving forward is, I believe, a path to easily get us to 20% plus growth.

However, I think there's actually a lot more upside from there. Scott, if you think about the adoption of generative AI, if you think about this space, particularly when it comes to the review piece of our space and the fact that it's a multibillion-dollar market that is being done by armies of human resources today, contract attorneys. We've got the ability to leverage our Auto Review capabilities on top of our core platform to turn much of that into AI-driven software revenue instead of services revenue.

And that is an incredible win for DISCO, obviously, but also for our -- the end customers, for the corporate clients who now will have the opportunity to bring the Ediscovery or the review process within Ediscovery much sooner in the litigation life cycle, which can help them improve outcomes. It can help them accomplish that part of the task for much more cost effectively. It will help our law firm customers increase their revenue streams to what's been traditionally done by these, again, armies of contract attorneys, and it will help us go along the way. So I'm actually optimistic beyond even the 20% plus growth profile.

Your next question comes from DJ Hynes with Canaccord.

Nice quarter. Nice to see the software acceleration continue, and I appreciate all the commentary on the call. Eric, maybe we could tackle kind of the elephant in the room. I mean you did a good job talking about kind of the moats that DISCO has kind of less directly hit on competition from the foundational model companies. Obviously, there's lots of noise in the market around those folks targeting legal tech is an attractive area for automation. Are you seeing those LLMs show up in your customers at all? Are they exploring with that technology? Can you talk about which areas of the tech stack are most at risk of disruption, which aren't and kind of how you think that impacts DISCO over the next 2 to 3 years?

Certainly, DJ. Sure, I happen to noted -- go ahead.

Sorry, I heard some feedback.

So yes, I did notice there was some disruption in the stock market recently, for sure, no question about it. But look, I've had the good fortune of spending time with our customers on a regular basis. Two weeks ago, I was in London in the U.K. and London and Manchester. 3 weeks ago, I was in Austin with our Customer Advisory Board members, and I haven't heard of a single customer utilizing general AI or these frontier models for the Ediscovery process. And frankly, I would have been shocked had I heard about it.

Look, it's a very different space. And I think it goes back to -- I look at it this way. You have to look at the industry that we're in specifically. You have to look at the competitive advantages that we have and then you have to look at the way we're innovating. The industry that we're in is squarely focused on litigation and Ediscovery and it's just a very different space than areas like contracts or M&A or transactional areas where these general AI and frontier AI companies are really focused.

And when it comes to litigation, either you win or you lose, and Ediscovery is at the heart of all of that. It's a complex. It's a court mandated. It's a legal process where the adversaries in a matter have to agree upon the methodology that they're using for Ediscovery. So they're dealing with extremely sensitive data that gets highly processed before it even comes into our platform or as part of coming into our platform, and it's really not valuable outside the context of the integrated workflow.

So -- and then on top of that, ultimately, lawyers can't make mistakes when it comes to litigation, it could cause a malpractice lawsuit or even worse yet, it could cause a crushing outcome for their firm's clients. So we're just in a very different segment of legal. Think of us as AI for litigators. That would be the first thing. And then as far as competitive advantages, look, we're -- DISCO, as I mentioned before, has been AI native since our inception, long before these frontier models came out. We were the first to embrace Gen AI in Ediscovery when we put Cecilia out 3 years ago.

And we've ultimately built a very powerful, scalable, integrated platform, as I detailed earlier, that just deals with the largest and most complex matters and with processed volumes of data that are well beyond what other solutions can handle. So I guess the way to think about it that way is that when it comes to technology for Ediscovery, that's really where DISCO is the answer.

And then I think you also have to think about the innovation. And DISCO is never sitting still. We've always been about innovation. Our entire history has been about improving the way litigation works. And as I mentioned earlier, if you think back to that story about one of our customers that in 2 days did what 70 contractors could do in 4 weeks, that's just a game-changing opportunity that can create a win-win-win for the corporate end client, for the law firm and for DISCO along the way. So I think we're in a better position than we've ever been, DJ.

Very helpful. Aaron, maybe a follow-up for you. You're obviously pretty fresh in the seat, but you've also a fresh set of eyes on the model and only the second CFO since the IPO, I'm curious of your impressions of the visibility that the usage-based model provides. And given this is your first call, maybe you could talk a little bit about how that informs your guidance philosophy?

Sure. When you think about the visibility the usage model gives us, I think there's elements that -- and I've seen this -- this is not my first experience with the usage model. I've had the privilege of kind of working with usage models in the past. But I think when you look at it, the larger the scale, the more predictable it becomes. You get to pick up the trends that occur within the model. And so I think as our business continues to scale, we pick up more and more predictability in the usage model.

There are still parts of the revenue model when you look at services and Auto Review as it stands today, where there is an element to that where it's less predictable. But once again, I think with scale, you gain the advantage of predictability, the law of large numbers type of math. And so I think going into our guidance philosophy, when you think about that part of it, we obviously provide it as a range for that reason. We know that our customers are voting with their wallets every time they choose to use DISCO. And I think that explains a lot of why we do provide the range. And it has -- if we look at Q1, our range in software is from 9% to 14%, and it's relatively wide for that reason. But I think as time goes on, it allows us to get more and more precise.

Our next question comes from Mark Schappel with Loop Capital Markets.

Nice job on the quarter. Eric, I just want to build on the earlier question about the new commercial model. I was wondering if you could just maybe discuss the Georgians a little bit more and maybe what potential downsides or trade-offs you foresee with this shift?

Yes. Thanks, Mark. Look, as Richard mentioned earlier, this really originated from our customers. If you think about it, our focus with our strategy on winning more wallet share within our biggest and best customers and getting large matters from their biggest and best customers, we've got great relationships with our champions at these customers. And sometimes, they have struggled to explain our pricing model to the various case teams within their firms. And so while we might be getting a nice $100,000 or $1 million worth of revenue from some of these customers, we might only have 15% or 20% of their wallet, and our champions want to do more with DISCO. They want to make sure that they can explain our pricing model.

Sometimes we've seemed more expensive than our competition when we're really not. And so we've had to try to teach our champions how to explain our pricing model to customers. And you can do that so long and you realize maybe there's an easier way to just simplify the model. I also think there's been a number of cases where we've had to discount more than I'd like to discount because our model wasn't as understandable. And so now with this new model, it's much more clear, much more understandable, and we think that gives us the chance to really proliferate along our strategy of getting larger matters and more wallet share within our largest customers.

Great. And then as a follow-up, the start of the year is typically when software companies adjust their sales organizations and their go-to-market strategies. 18 months, 24 months or so back, you made a significant change to the sales org. I was wondering if you could talk about any meaningful changes to the sales org as we start the year here?

Yes. Thanks, Mark. We've had the -- the strategy that we execute upon with our go-to-market shift has worked really, really well. And it started with ensuring that we're bringing in the right leaders, the right talent, the right reps. It was a matter of -- as you know, we made some shifts last year moving from less account executives to more outside salespeople, so more from inside sales to more outside sales as we're focusing on larger customers and larger matters. So we didn't add cost to sales and marketing last year, but we did shift the way we spent that money, and it's paid off.

Also the comp plan that we put in place to really incentivize sales reps for new matters and new revenue has worked out really well for us. The systems and processes that we put in place, the contract simplification. So a lot of the things that we have already put into place are starting to work. And so the main thing that we're doing is just doubling down and executing on that.

Now as I mentioned, we didn't add a lot of cost to sales and marketing in 2025 because we needed to go through that process of sort of reorganizing and reaccelerating revenue. I think there's an opportunity this year potentially to bring in some additional talent to take advantage of the opportunity that we see ahead.

There are no further questions at this time. I'll now turn the call back over to DISCO's CEO, Eric Friedrichsen, for any closing remarks.

Yes. Thank you very much. Look, to wrap up, our performance in 2025 was remarkable. It gave me extreme confidence that our strategy focused on expanding our wallet share with existing customers, focusing on large and strategic matters and accelerating our Gen AI adoption of Cecilia Auto Review on the right strategy that focusing on our customers and with you in every case, combined with the innovation that we're delivering has really put DISCO on the right path.

So we're going to do a lot of the same things that we did in 2025 and '26. We're going to take advantage of the momentum that we started to gain. We're going to layer on top of that the new Agentic AI capabilities that we just announced and our new pricing and platform approach. And so look, it's been 3 years now where you've been able to see the acceleration, I should say, 2 years on top of 2023, where you've been able to see the acceleration. I think it's a very large market with an Ediscovery and litigation that DISCO is in a prime position to go disrupt. And I'm really excited. I think it's going to be a great year. So I look forward to updating you as we progress throughout 2026, and I really appreciate you all joining. Thank you.

Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.

Ladies and gentlemen, thank you for standing by, and welcome to CS Disco's Third Quarter of Fiscal Year 2025 Conference Call. [Operator Instructions]

I would now like to turn the call over to your first speaker today, Head of Investor Relations, Aleksey Lakchakov. Alexey, please go ahead.

Good afternoon, and thank you for joining us on today's conference call to discuss the financial results for DISCO's third quarter of fiscal year 2025. With me on today's call are Eric Friedrichsen, DISCO's Chief Executive Officer; Michael Lafair, DISCO's Chief Financial Officer; and Richard Crum, DISCO's Chief Product, Technology and Strategy Officer.

Today's call will include forward-looking statements within the meaning of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995, including, but not limited to, statements regarding our financial outlook and future performance, our future capital expenditures, market opportunity, market position, product and go-to-market strategy and growth opportunities and the benefits of our product offerings and developments in the legal technology industry.

In addition to our prepared remarks, our earnings press release, SEC filings and a replay of today's call can be found on our Investor Relations website at ir.csdisco.com.

Forward-looking statements involve known and unknown risks and uncertainties that may cause our actual results, performance or achievements to be materially different from those expressed or implied by the forward-looking statements. Forward-looking statements represent our management's beliefs and assumptions only as of the date made.

Information on factors that could affect the company's financial results is included in its filings with the SEC from time to time, including the section titled Risk Factors in the company's annual report on Form 10-K for the year ended December 31, 2024, filed with the SEC on February 20, 2025, and the company's upcoming quarterly report on Form 10-Q for the quarter ended September 30, 2025.

In addition, during today's call, we will discuss non-GAAP financial measures. These non-GAAP financial measures are in addition to and not a substitute for or superior to measures of financial performance prepared in accordance with GAAP. Reconciliation between GAAP and non-GAAP financial measures and a discussion of the limitations of using non-GAAP measures versus our closest GAAP equivalent is available in our earnings release.

And with that, I'd like to turn the call over to Eric.

Good afternoon, everyone. I am delighted to be here with you all today to report on another quarter of accelerated revenue growth and operational execution for DISCO. Software revenue in Q3 was $35.2 million, up 17% year-over-year, while total revenue in Q3 was $40.9 million, up 13% year-over-year. Adjusted EBITDA for Q3 was negative $297,000, representing an adjusted EBITDA margin of negative 1%, which is a $4.2 million improvement over Q3 of 2024.

Our results include $1.3 million of revenue related to a matter that was contingent on a successful outcome of the case. Michael will speak more to this later. But even without this bump, I am pleased to share that we exceeded the high end of our guidance range for software revenue, total revenue and adjusted EBITDA.

We finished the quarter with $113.5 million of cash and short-term investments and no debt. We ended Q3 with 326 customers who each contributed more than $100,000 in total revenue over the last 12 months. The proportion of revenue attributable to these customers is 76%. We're extremely excited about our performance this quarter. The growth acceleration coupled with our improved operational execution is a testament to our larger strategy.

Last quarter, I discussed our focus of the growth of large multi-terabyte matters, and we've seen continued positive trends this quarter in that area. The usage patterns in our business have continued to give me optimism. This quarter, we saw continued growth in revenue from both large and small matters with acceleration in revenue from multi-terabyte matters.

We are also very pleased with the continued adoption that we are seeing related to our suite of generative AI capabilities, including Cecilia AI and our AI-driven Auto Review. The number of customers utilizing Cecilia AI over the quarter more than tripled year-over-year, and we have experienced consistent sequential growth in Auto Review adoption throughout 2025.

Our overall positive performance continues to be driven by our relentless focus on delivering value to our customers through our value proposition of with you in every case. A great example of this is with a large multinational company involved in an industry-transforming lawsuit. They selected DISCO as their legal technology provider and services partner. The case involves more than 10 terabytes of data, and this customer chose DISCO because we demonstrated our end-to-end capabilities in handling very complex cases.

This customer selected DISCO because they value the speed, scale and ease of use of the DISCO platform. They were able to reduce time to evidence while searching across hundreds of thousands of conversations, spanning multiple communication forms and time periods. And in terms of support, they valued our onboarding and our in-house services capabilities. They appreciated their forensics team was on call to handle dozens of data sources, that we had a fully staffed project management team and that we had leading review experts who could partner with them to deliver end results at an exceptional speed.

This customer was sourced through our lead generation team, leveraging the new territory-based account orchestration model that I mentioned last quarter. This is just one example of that change already yielding results. The entire sales process for this customer was completed in roughly 4 months, a clear demonstration of the strength of our strategy and execution.

Our results this year reflect a sharper focus on the right customers. Over the past year, we refined our approach to target those that we believe are the best fit for DISCO based on their scale, industry and complexity as well as the specific matter types to best take advantage of the capabilities of the DISCO platform.

This is further validated by IDC, who named DISCO a leader in the 2025 IDC MarketScape for worldwide end-to-end eDiscovery software. The IDC MarketScape noted that with the ability to scale seamlessly from modest projects to multimillion document matters, DISCO's tools handle large data volumes effectively, facilitate thorough analytics and support core workflows such as time line creation, deposition summarization and intelligent batching, all while maintaining accuracy and transparency for defensibility. This is a powerful statement by IDC that really encompasses what we strive to do here at DISCO.

As you know, we have previously discussed how our platform is particularly beneficial for certain types of matters. One of these matter types is intellectual property litigation. This past quarter, we began a new initiative that highlights our strength in IP litigation over our competitors, focusing a portion of our marketing and sales efforts specifically on IP litigation.

Customers tell us that DISCO is an ideal fit for IP litigation for 3 key reasons. First, IP litigation cases are typically very large and highly technical, involving decades of research, product development files, technical specifications, e-mails, source code, CAD files and other complex document types. DISCO's ability to handle complex file types and deliver at scale makes us a key enabler of client success, and we shine brightest when building deep trusted partnerships in these types of environments.

The second reason is that these matters tend to be high stakes bet the company litigation. A company's entire business plan may hinge on the outcome. DISCO's AI embedded within our platform helps lawyers quickly identify and understand the most relevant materials seamlessly across diverse file types, providing a strategic case-defining information advantage to the legal teams.

Third, these cases typically involve aggressive time lines set by the courts and sheer volume of work that makes speed and precision critical. Our AI-assisted workflows with Cecilia AI and Auto Review significantly reduce time to evidence, enabling our customers to hit deadlines and develop an optimal case strategy without sacrificing accuracy or quality.

Our unique capabilities make DISCO the natural solution for IP litigation that can deliver significant value for our customers while driving long-term growth for DISCO. As a matter of fact, this customer success story that I mentioned earlier is an IP-related matter that perfectly fits our strategy. In the future, you will see us roll out similar initiatives for other matter profiles where DISCO delivers immense value compared to our competitors.

Finally, we are operating in a highly complex environment. The legal world is changing rapidly, and DISCO is in a prime position to be the disruptor in this technology revolution. We have a core platform that makes complex workflows look simple and effortless, which is then paired with AI capabilities that are transformational to how our customers approach their craft.

I often get questions from you about Cecilia and our generative AI innovation and how DISCO is different. With that in mind, I want to take some time for you to hear directly from Richard Crum, our Chief Product, Technology and Strategy Officer, on this specific topic, and I'm excited for you to hear from him.

So with that, I'll turn it over to Richard.

Thanks, Eric. I appreciate the opportunity to share more about how and why our technology is a strategic advantage in this really exciting time. I joined DISCO 16 months ago, knowing of our reputation for having a strong product and for leveraging modern cloud technology and AI to offer a solution that is intuitive, innovative and operating at scale with impressive performance metrics.

I also joined looking forward to teaming up again with Eric, whom I worked with as Chief Product Officer at Emburse. Here at DISCO, I have the privilege of leading our product and engineering teams, and I've seen close up how effective we are at offering solutions that meet the high bar our customers have for the technology they use to achieve the best outcomes on the legal matters they manage.

Let me give you some examples of why that is true. It shouldn't surprise you that I want to start to tell you about the AI capabilities that make our products so effective with customers. Before I talk about the GenAI features that understandably get more airtime, I think it is important to note that DISCO has been building AI throughout our products for over a decade in ways that directly impact how our customers manage legal matters.

Our dynamic topic clustering technology is often the first place a user of DISCO will go after ingesting all of the likely relevant documents. The analytical capability of the models that powers this set of features gives attorneys a quick glance into the people, entities, information and topics resident in the document population.

The technology automatically updates the information as new data and documents are added to the platform. Attorneys use this tool set to help understand and analyze thousands or millions of documents in real time. And they can do this without requiring the help of a services team or needing to leverage a separate solution. It's one of the many great examples of the power of the DISCO platform and something we know drives real value for our customers.

A second example is our predictive analytics that observes the work attorneys do in DISCO and directs them to the other documents and data that are likely to be highly relevant based on how they have interacted with other documents. When you have a database with terabytes of information, the time savings this capability offers can be massive. And helping legal teams quickly narrow down the population of documents down to the right set enables them to focus on the evidence that matters most.

In the last few years, we have built on the foundation of these examples of AI-powered tools to offer additional GenAI features that we call Cecilia. They further enable legal teams to identify relevant evidence quickly and with confidence. Cecilia Q&A operates at the level of an individual document or across an entire database that could contain millions of documents.

In a simple context-aware chatbot, lawyers can ask natural language questions and interact with the data as if they were speaking to an associate who has read and fully understood every single document and get answers with reference only to the evidence contained in the documents because that's what we built Cecilia to be. And it's not just impressive at finding the key documents quickly and explaining to you why they're relevant to your inquiry. Cecilia is built to be a trusted tool for attorneys.

In a world that worries about generative AI hallucinations and what data a model was trained using, we developed Cecilia to be technology that customers can have confidence in. When it returns results, it also provides citations back to the exact part of the document that is used to answer your question. And its answers are only based on the documents in your database.

This is not a simple LLM wrapper. We develop Cecilia with a set of technical sequences that deliver a product experience that is powerful and impactful to your legal work. Cecilia Q&A is just one of the impressive GenAI skills that is built right into the DISCO platform at moments in the workflow that matter. We also offer document summaries, definitions of any term based on the information of the document set, automatic time lines and topic summarization of deposition transcripts.

Altogether, Cecilia is a sophisticated set of tools, leveraging generative AI technology that we have been offering to the market for almost 2 years and constantly investing to make it better. Tools built upon the existing system of record the legal team is already using to manage the matter. And when legal professionals experience the impact, it can be hard to imagine working on a matter without Cecilia.

A great example comes from an Am Law 50 customer that tried out Cecilia AI for the first time in July of 2024. They saw the value that is provided by utilizing these capabilities I just described. And since that initial use of Cecilia, their adoption has continued to expand. Following their initial trial, this customer has adopted Cecilia AI in a variety of types and sizes of matters from matters as small as 5 gigabytes to large complex matters with millions of documents.

In fact, the number of matters utilizing Cecilia AI at this firm has grown 7x from Q3 2024 to Q3 2025, which corresponds to a more than 12x growth in revenue. We are proud to continue partnering with this innovative customer, delivering advanced generative AI solutions that turn time-consuming legal challenges into streamlined, high-value opportunities.

Another DISCO GenAI capability that has been delivering significant value to customers for more than a year now is Auto Review. You have heard Eric and Michael talk about Auto Review since it was launched last August and how impressed customers have been with its performance, accuracy and the cost savings it offers, particularly on very large matters.

From an engineering standpoint, it represents something much more sophisticated and technically challenging than sending off documents to a large language model. The technology behind Auto Review is built upon a decade of deep understanding of the review process and technology that was developed to handle the scale and complexity of documents and data that our customers bring to DISCO.

This technology earns our customers' trust by providing clear explanation for tagging performed by Auto Review and offering tools to report quality statistics that align with how courts measure the efficacy of other technology-assisted document review approaches.

I've spoken a few times about trust when describing our AI-based capabilities. We take that very seriously, and it extends even back to the core technology platform where we bring all these features and capabilities together in a simple yet powerful user experience. At DISCO, we obsess about ensuring the entire product experience for our customers is secure, reliable and incredibly performant at the scale of the matters that our customers bring.

The volume and sources of data that could be necessary to review for evidence and litigations, investigations and regulatory matters has continued to explode and no one believes there's an end in sight. This has resulted in customers bringing larger and larger matters to DISCO, and we have continued to perform with the speed and accuracy they've come to expect. That's because DISCO was built for scale. Leveraging the best technology and engineering talent, DISCO is ready for whatever our customers need.

Let me finish up with a word on another phrase I've used more than a few times in my commentary, and that's scale. Because of our personal experience with consumer technology that can feel like it's improving at an astronomical rate, we can begin to believe that business software in general is also advancing at the same rate, but that's not always true. In fact, much of the legacy software used for document review is far behind DISCO and how well it performs under high database sizes.

Ask any legal professional about their experience in legacy Ediscovery solutions with simple tasks like switching between 2 documents or running a keyword search, and you're very likely to hear the word slow in their response. That's because operating at scale doesn't just mean being able to hold a large quantity of data. It also means providing a user experience that runs just as well in both small and massive sets of data and documents.

This is an area where DISCO shines. And we do it both with those simple tasks I just referenced and the experience of using Cecilia Q&A, something that definitely leaves our users amazed when they see it live. Operating a platform that performs like DISCO is the result of years of solid engineering work and our obsession with providing an incredible user experience to everyone who logs into DISCO every day. And it is a great example of how we are with you in every case.

It is not easy to operate AI at the scale the way we do at DISCO. It is also a competitive advantage that will enable DISCO to continue to expand our product offering and win more and more loyal customers who say it has to be DISCO.

And with that, let me turn things over to Michael.

Thank you, Richard. In Q3 2025, total revenues were $40.9 million, up 13% year-over-year, and software revenues were $35.2 million, up 17% year-over-year. Included in these balances is the revenue contributed from a case that has been on our platform for several years and was contingent on the successful outcome of the case. In Q3, we were able to recognize $1.3 million of total revenues from this case, of which $1.2 million related to software.

I would like to note that the Q3 total revenue and software revenue year-over-year growth would have been 9% and 13%, respectively, without this contingent revenue, which still exceeds the high end of our guidance range for both metrics.

Excluding the large one-time case we recognized in Q3, 2 primary drivers of the software year-over-year revenue performance were growth in the revenue across large and small matters, especially with multi-terabyte matters as well as growth of Cecilia AI adoption. Services revenues, which include DISCO managed review and professional services, were $5.7 million.

In discussing the remainder of the income statement, please note that unless otherwise specified, all references to our gross margin, operating expenses and net loss are on a non-GAAP basis. Adjusted EBITDA is also a non-GAAP financial measure.

Our gross margin in Q3 was 77% compared to 74% in the prior year. As we mentioned before, our gross margins fluctuate from period to period based on the nature of our customers' usage, for example, the amount and types of data ingested and managed on our platform.

Sales and marketing expense for Q3 was $13.6 million or 33% of revenue compared to 38% of revenue in Q3 of the prior year. On a dollar basis, sales and marketing expense decreased $0.2 million, predominantly driven by lower marketing and consulting expenses.

Research and development expense for Q3 was $11.5 million or 28% of revenue compared to 31% of revenue in Q3 of the prior year. On a dollar basis, research and development expense increased $0.4 million, driven primarily by headcount-related costs. General and administrative expense in Q3 was $7.7 million or 19% of revenue compared to 21% of revenue in Q3 of the prior year.

Adjusted EBITDA was negative $0.3 million in Q3, representing an adjusted EBITDA margin of negative 1% compared to an adjusted EBITDA margin of negative 12% in Q3 of the prior year. This represents a beat of the high end of the guidance range we provided last quarter and $4.2 million year-over-year improvement.

Net loss in Q3 was $0.6 million or negative 1% of revenue compared to a net loss of $3.9 million or negative 11% of revenue in Q3 of the prior year. Net loss per share for Q3 was $0.01 compared to $0.06 per share for Q3 of the prior year.

Turning to the balance sheet and cash flow statement. We ended Q3 with $113.5 million in cash and short-term investments and no debt. Operating cash flow for the first 3 quarters of 2025 was negative $15.7 million compared to negative $10.8 million in the same period of the prior year.

Turning to the outlook. For Q4 2025, we are providing total revenue guidance in the range of $38.75 million to $40.75 million and software revenue guidance in the range of $33.75 million to $34.75 million. We expect adjusted EBITDA to be in the range of negative $3.5 million to negative $1.5 million.

For fiscal year 2025, we anticipate total revenue guidance in the range of $154.4 million to $156.4 million and software revenue guidance in the range of $132.6 million to $133.6 million. I would like to note that the contingent revenue case I spoke about earlier was previously included in our full year guide. We expect adjusted EBITDA to be in the range of negative $11.5 million to negative $9.5 million.

Now I'd like to turn the call over to the operator to open up the line for Q&A. Operator?

[Operator Instructions] And it appears there are no questions. So I will now turn the call back over to CEO, Eric Friedrichsen, for closing remarks. Eric? Actually, Eric, we did just receive one question, if you would like to take it.

Sure.

Okay. Our first question comes from the line of DJ Hynes with Canaccord.

I guess I just wanted to ask about the contingent liability or the contingent case rather. How many cases are like that? Is that industry standard or it's still a little odd?

DJ, let me touch on that real quick. We have a small number of other contingent cases in the system, but nothing close to this size. Let me explain by what we mean by a contingent case. It's basically on a limited basis, we'll enter into a contract where basically the payment is contingent on the conclusion of the legal matter. And in these instances, we don't recognize the revenue until the legal matter is resolved.

In Q3, one of our customers under one of these arrangements experienced a favorable conclusion, allowing us to recognize $1.3 million in revenue, of which $1.2 million was software and the balance was services.

Yes. And so that's why we called it out specifically, DJ. We had 17% software growth and 13% overall growth. But even without that case, we had 13% software growth and 9% overall growth in the business, which exceeds the high end of our range even without that contingent case, but we did want to make sure to call out that contingent case specifically.

Okay, great. And then just a quick one for Michael. So we have your previous guidance or I guess your target for in-quarter EBITDA breakeven for Q4 of 2026. But I'm looking at -- we're seeing some linear improvement over the past couple of quarters where you were just under breakeven this quarter. Are you still maintaining that target for next year or do you think there's some upside we could see there?

I'll take that one. This is Eric. Yes, we're still targeting adjusted EBITDA breakeven for Q4 of 2026. We could certainly push to get to profitability sooner. But right now, we're making really smart investments in the business that are obviously paying off in the results in terms of the go-to-market investments that we make, in terms of the innovation investments that we make and those transformational investments. So look, this is a big market. It's a growing market, and the money that we're spending right now is really helping accelerate the revenue of the business. So the target is still Q4 of 2026 for adjusted EBITDA breakeven.

[Operator Instructions] All right. Do not see any callers. So again, I will hand the call back to CEO, Eric Friedrichsen, for closing remarks. Eric, take it away.

Great. Thank you very much. Thanks, everyone, for joining the call today. Look, I am really pleased that our progress in Q3 and really, frankly, our progress throughout the entire year. We've been very fortunate. We set a strategy for this year to focus on our biggest and best customers with the most opportunity for growth and also the matter types where we could add the most value to our customers and also generate the most revenue for DISCO.

So we set that strategy, but more importantly, I'm really proud of the team for how they've executed upon that strategy. And it's given us the opportunity in each of the quarters this year to be able to beat our guidance -- this last quarter to be able to beat the high end of our guidance range and for us to be able to increase our guide for the full year every single quarter.

But it's not even so much the results that I'm happy about. It's the team and how they've executed upon our strategy. And that gives me a lot of confidence. It gives me a lot of confidence about our future. So I just want to thank our teams. I want to thank our customers for everything they're doing to stay focused on accelerating the growth for DISCO. And I'm looking forward to updating you next quarter. Have a great evening.

Thanks, Eric. And this concludes today's conference call. You may now disconnect. Have a great day, everyone.