Teleperformance SE Aktie

Welcome to TP First Quarter 2026 Revenue Conference Call. [Operator Instructions] Now I will hand the conference over to Jorge Amar, CEO. Please go ahead.

Excellent. Good morning, good afternoon, good evening from wherever you are joining us, and thank you for being with us in the TP Q1 2026 Revenue Update Call. My name is Jorge Amar. And as of March 16, I have the honor of being the CEO of the TP Group. Today, I'm joined by Benoit Gabelle, our Interim Chief Financial Officer and he will walk us in a bit through the more detailed numbers for today.

Before we begin, I just want to say that I look forward to these communications with the investor community. I've had the pleasure of talking to many of you over my first few weeks in the role and your conversations are always very energizing. They keep us very honest. They help us sharpen our perspectives. So thank you for your engagement. Thank you for your perspective, and thank you for the questions that I'm sure will be very vigorous as always.

So with that, we have a very structured agenda for today. I will walk us quickly through some of the key highlights of the Q1. I will then hand it over to Benoit for a more detailed update and the walk through the revenue numbers. And then I will just share with you some of the interesting things that I've been hearing by talking to our clients, our partners, our employees, the investor community and just go through that in a little bit more detail. So you can know what to expect for the rest of 2026. After that, I will open it up for questions, as always. And I'm sure we already know that you have a few and we are ready to take them.

So with that, let me get us into the quick -- the key highlights of Q1. So we start with as we had anticipated in our last series of conversations mostly after our annual results that we were expecting a softer Q1, which is exactly what has played out. So the overall revenue of the group came down at EUR 2.433 billion that is a minus 2.2% in a like-for-like basis, minus 6.9% as reported. What you see there from a Core Services perspective is that, indeed, the revenue came down at minus 1.7% like-for-like, reported minus 6.4%.

What we are seeing on the Core Services on one side is a number of our vertical, both from a back office perspective, data labeling, data services perspective, that continues with tremendous double-digit growth and you will see a little bit more in detail in a few minutes. On the other side, the headwinds that we are experiencing are mostly 3. So let me describe them a little bit more in detail for you.

The first one is we see and we experience increased offshoring, which in many ways, translates in good growth for some of our delivery locations, but of course, with the corresponding pressure on the top line at an overall level. We also continue to see the trend that we had started experiencing a few quarters ago on the automation of trust and safety. And we continuously monitor this space as we have seen some developments mostly in the U.S. for this particular line of business.

And last but not least, we have seen that the current geopolitical environment has delayed the ramp-up of some of the contracts that we were expecting in Q1. So the combination of those 3 things are driving the results that we report today. We know them. We don't minimize them, but we know exactly what they are, and we are working for each of them in more detail.

When it comes to specialized services, revenue for Q1 is down minus 5.5%, reported is minus 9.9%. It's around EUR 332 million. And on this case, in particular, what we see is two effects. The first one is we continue to operate against a very high basis of comparison for LLS considering that some of the elements that have impacted this particular company of the group started developing at the end of Q1 2025 in the U.S.

On the other side, this is the last quarter that we will see the high basis of comparison for TLScontact, our visa processing business due to the loss of a contract that we experienced around Q1 last year. So those are the two elements that are explaining the softness in specialized services.

I will get into a little bit more detail later on what we are seeing also not only on the numbers from a revenue perspective, but the momentum that we're getting in the acceleration of our 2026 plan for the Future Forward transformation. We're having a lot of very interesting conversations with clients that are looking for a reliable transformative partner into their operations, and I will elaborate a little bit more into that.

We continue to see great traction of our TP.ai FAB solutions, mostly in some of the new deals that we are getting pretty much every deal that we are winning today has a component of this, and I'll expand that in a little bit more detail later. We have also made some changes in our AI leadership. So we are very happy to announce that we have a new Chief AI Officer for the group. His name is Andreas Braun, and I will describe his profile later on. And we're also reporting that our efficiencies program that we had already reported and highlighted to the investor community continues to be very well on track which leads us to confirming our 2026 targets. This is just a quick summary. We will get into more details now. So I'll pass it on to Benoit, so he can walk us through the revenue breakdown in the next few slides. Benoit?

Thank you, Jorge, and good morning, good afternoon and good evening to everyone. We'll now look into more detail on the Q1 '26 revenue. You will see in the press release the detail of the numbers that we are showing now. Jorge, you mentioned the reported decrease of minus 6.9% overall for the group and a like-for-like negative growth of 2.2%. I just want to give some details about the impact of the acceleration of our TP Future Forward plan, where we have decided to combine all our collection business activities from core plus AllianceOne into one single management to accelerate on synergies and business efficiencies. So that leads us to transferring the AllianceOne revenues from specialized services to Americas region starting this year.

Smaller impact, TP Infinity was mostly present in Europe in the past, so was reported in EMEA and APAC, and we have seen a growth in the U.S. in the past year. We expect still a good growth for this year. And so we have taken a stance to report the Americas part of TP Infinity into Americas region.

Now moving more on the detail of the absolute numbers. So we had a revenue of EUR 2.6 billion in Q1 '25. As you mentioned, Jorge, the biggest part of the decrease that we have observed in Q1 is linked with the effects with an impact of EUR 141 million on our top line. Now talking more about the like-for-like comparison, EUR 55 million less of revenues, minus 2.2%, which includes almost nonmaterial hyperinflation impact of plus 0.1% linked with our operations in Argentina and in Turkey, and we have a perimeter effect linked with the entry into the consolidation of ZP, which was reported only for 2 months in Q1 2025 and is reported 3 months this year and Agents Only, a business that we acquired in June '25 and which has reported revenues this quarter, whereas it was not present in the group in Q1 2025.

Now moving on to the detail of the revenue by activities. You see that the balance in terms of revenues between Americas, EMEA region and Specialized Services have not moved very significantly compared to last year. We have 2% less for Americas balancing with Specialized Services in EMEA and APAC. This is also linked with the impact of FX because Americas is mostly reported in USD. The revenue for Core Services, minus 1.7% like-for-like. This is hiding, unfortunately, the good news that you mentioned, Jorge, and acceleration of our back office and AI-powered solutions offering with good growth. And this was effectively hidden because of the soft start of the year linked with, like you said, offshore delivery, big increase of our activity in India, in Egypt, in South Africa, which are already very large operations, but still growing very fast.

We also see the impact of trust and safety, which is automated as we expected or as we anticipated, which is being automated and which we are closely, rebalanced, like you said, with our AI offering, but not completely offset yet. And a slower-than-expected ramp-up for some of the large contracts linked with the geopolitical turmoil and some clients taking more time for decisions.

Specialized Services, minus 5.5% like-for-like. Jorge, you mentioned that Q1 '25 was the last time when the TLS contract that was lost was still in effect. Restated for that impact, we would be at minus 1% like-for-like. Still, this is not in line with what we would like to see and the big difference is linked with the high comparison basis for LLS in Q1 2025. But we have seen, and that was a good news especially in March, a good sequential improvement of the revenue growth throughout the quarter, which is giving good hopes for the coming month, depending on how the policy in the U.S. will evolve.

Now if we look more at the breakdown by business line, we still see a good growth in Care, which is most of our Core Services business. We have seen, like I said earlier, very strong traction on AI-powered solutions on back office and also to some extent, on sales, which we know is a good business in terms of margin and reliance over time I will not repeat again the impact of the trust and safety, which has moved from 8% of the group revenue last year in Q1 to 6% this year.

Moving on to the view by vertical. What we like and I think is a key differentiator of TP compared to some smaller competitors is the very diversified portfolio that we have with clients operating from almost all major verticals. This was particularly true, I mean, in terms of benefits to be present, notably in financial services and insurance as well as FMCG for Q1 2026, where we have experienced double-digit growth and that we see continuing during the year. This helps offsetting the lower activity in the automotive industry, energy and utilities and the media and entertainment. Media and entertainment being the vertical where you will find most of the trust and safety business.

With that, Jorge, I hand over to you for an outlook on the months to come.

Excellent. Thank you, Benoit. And as, of course, you see and you can feel from our tone, we're definitely not where we want to be in terms of the revenue that we are reporting to you, but this was anticipated. And this was already what we were seeing and what we had communicated to the investor community in the last time that we were able to chat, which was just a few weeks ago as I was being -- as I was appointed the CEO. But with this, let me also turn on the other side, what to expect for the rest of 2026. And as I continue my tour of talking to, again, a number of our clients, our partners, our employees, start sharing back with you what I'm hearing and how that can inform some of the movements and some of the actions that we will take in 2026.

I think this one is just a quick reminder, and I think probably Thomas and Olivier shared a version of this in their last set of communications around the direction of travel as part of our Future Forward strategy, which has a component around how do we reimagine and evolve our core service, mostly because what we are hearing from clients is the need of a partner that is a stable partner that they can trust on and that they can work on some of their own bigger, more transformational ideas.

So that's a big component and that we continue working on as we also see that many of our partners are starting and have started the discussions around vendor consolidation and TP is at the center of that conversation as well, and we're very well positioned in that space.

The second pillar is around all the work that we are doing to migrate more and more of our revenue to an outcome-based arrangement. We continue to see very strong performance and growth in our sales and revenue as a service line of business, the same we're seeing in collections, the same we're seeing in a number of vertical back-office solutions. And all that gives us a lot of hope of the potential that those lines of businesses have.

And last but not least, AI is also opening up a new series of opportunities, LLM and SLM model training, how do we do more data labeling, data annotation. So how do we become much more relevant in that space that it's opening up a new set of opportunities in what I call the AI value chain. Of course, doing that, it doesn't happen by accident. This means that we need to take a number of actions when it comes to reshaping our operating model, our processes, our talent, our skills and capabilities, our partnerships. And all that is what you will start seeing more and more, and I look forward to sharing more of the details as I go through this transformation.

I've also taken the time to talk to many of our top clients and thank you for the trust, the candid feedback and the very open conversation on that. What I've heard is on one side, a number of our clients are in this transformational moment. They are thinking about how to leverage AI and other technologies into their own operation. And to do that, they need that reliable partner. They need that partner that knows the inner workings of many of these organizations and know their processes, their data. And that's what makes us being very, very well positioned, uniquely positioned in that space. A number of our clients are coming to us not only with a conversation about what I call, efficiency AI, which is doing exactly what you're doing today with fewer resources.

A number of the conversations that I've had are around what I call opportunity AI, which is everything that you wish you could do that you're not doing today because of cost, because of limitations, because of time to process and we're engaging in those conversations. And that's incredibly exciting because of the opportunities that we can unlock together is just the opportunity is massive.

I've also heard a real concern from a number of the leaders in our space about what is the cost of technology. One thing is to execute a small proof of concept with just a few things here and there. But when you start scaling that up and when you start looking at what is the overall profile of cost, that remains a point of concern as they are seeing how quickly some of those costs can increase and vis-a-vis some of the human activities, how to balance it all.

And last but not least, from a technology perspective, we continue to see the emergence of a number of different AI players with very unique and very distinct capabilities. I think we're also encouraged by some of the developments in this space, considering just how easy it has become for many companies to develop some of these solutions and the same for us that we're partnering with our clients in doing that. And therefore, we believe that being what I call tech agnostic, but in a way, means being able to quickly react and be very agile in partnering with our clients on their technology solutions remains something that we are focused on.

I already teased it out that we are hearing more and more this notion of being more of a transformational partner that is just not only delivering headcount, but is delivering a true re-imagination of a process, a true reinvention of how to treat customers, how to handle some of the concerns that our customers have and they are turning to us, they're outsourcing partners to be a key element in that. And that is, of course, something that you cannot do with a multitude of vendors and we are hearing more and more, and I've been in many conversations on this around how to drive vendor consolidation. And therefore, they want to be with a partner that is not only reliable that they can trust that they have been with. As you know, our average tenure of relationship is over 13 years. But they want to be also with a partner that has the financial strength and stability to be that partner long term.

So that is all that I've heard from our clients and something that will definitely inform more and more the type of actions that we drive internally. On the other side, I've already made some decisions when it comes to aligning our business and getting it closer to our clients. Our management teams are fully committed to delivering that and to accelerate the transformation, and we are seeing very, very promising early results that I really look forward to sharing more with you in the coming calls.

We have also made some changes to our IT and AI operating model. And as such, we are very happy to announce the appointment of Andreas Braun as our new Chief AI Officer for the group, that he will be joining us as of Monday. And so on that, just a quick few words on Andreas. Andreas, we're very, very happy he's joining us, and he will own our mission to scale our AI capabilities, scaling them from a client-facing perspective, scaling them internally as we turn some of the internal things that we're doing also into some business opportunities for us and to work with that partner ecosystem that we have. And thank you for all the multiple AI companies that have reached out over the last few weeks to talk and to have -- and to share ideas with us.

Andreas comes most recently from BCG X, Boston Consulting Group. Before that, he was with Microsoft. Before that, Allianz, before that with Accenture. And I think his profile is very unique, is very unique because he's not approaching AI only from a technology perspective, he also approached it from a business building perspective. So that is what gets us really, really excited and we really look forward to working with him over the next few years.

On top of Andreas, we will also have a team of very experienced practitioners and AI professionals that are joining us as well in the next few weeks, and we will share more with you as we finalize the revamp of this organization.

Now what are they inheriting? Well, they are inheriting a very robust set of solutions that we have been able to create in TP.ai FAB. We are very happy to announce and to report that we're roughly -- we have over 550 client AI projects live today in Q1 2026. You can see there that we have seen an acceleration in the momentum in Q1 with over 50-plus deals that we have won in this space, and this gives us a lot of hope in terms of we see the traction.

This is not just a technology marker, this is a real business driver because we are being able to win business, we're being able to secure business from our clients given the strength of some of these solutions. So maybe let me bring that to reality and to life a little bit more. And here, we have an example, and this is a client. It's a large financial institution that we serve in India and they came to us with a clear exam question of like, hey, help me manage my customer interactions. And help me manage them with your human workforce, help me manage it with your hybrid workforce, so TP.ai FAB Connect, our Agentic [ AI ] solution.

So we were able to deeply partner with them. And what gets me really excited and why I wanted to highlight this example with all of you is we're not only a tech provider, we're not only a human provider, we're bringing the interconnection and the feedback loop between the Agentic workforce and the human workforce. On top of that, we were able to partner with them into embedding the solutions from just not only the contact center technology, but also getting into their core banking system and their architecture. And with that, we were also able to bring some of our expertise on back-office solutions for banking and financial services, which is starting to become much more relevant. In a world where the technology is becoming more and more a commodity, the expertise and not just the -- I say I had expertise, but the real expertise of process excellence, process reengineering, understanding the limitations on the system, understanding the functionalities and everything that can be possible for our clients becomes essential.

So therefore, this solution, being able to, again, bring the Agentic workforce, the human workforce, the deep understanding and expertise in this particular vertical allowed us to not only secure this client, but also drive some vendor consolidation in this space, which is what gets me really, really excited about it. And that's why I also wanted to share just to bring to life that some of the numbers that we see in this slide are translating into these business opportunities.

So with that, what you see here is our commitment to the 2026 guidance on a group like-for-like revenue of roughly 0% to 2% even in the current market context, the same delivering the guidance on stable margin at 14.6%. The same on the free cash flow of EUR 800 million to EUR 850 million excluding, of course, nonrecurring items that we have highlighted already, and you see below in that slide of roughly EUR 70 million to EUR 90 million of restructuring costs expected in 2026. And we see a softer H1 but a true acceleration on the second half of the year, given the pipeline that we are seeing, given the conversations we're having with clients, given the momentum that we have seen in our Specialized Services division that continues to accelerate month-over-month. So all this gives us the confidence on these numbers.

And last, I briefly touched it at the beginning of our call, we are also very pleased with the development of our Future Forward efficiency program that is not only driving efficiencies from just a pure procurement space, for instance, but that is really helping us retool how we do work, how we go to clients, how do we contract, how do we price. A number of things that I've heard very loud and clear that our clients want us to make progress on. And that is -- we're seeing the performance, we're seeing the delivery, we're seeing the actions happen.

So all of these for 2026 and then, of course, as we have already communicated our long-term guidance that you see here in this slide that we've covered many, many times before.

So with that, and before we open it up for questions that I'm sure you have many and we'll be here to answer them with Benoit. Again, I don't want to say that we're happy with the results, but this is what we anticipated coming in. But hopefully, what you take away from this is that I'm now on the ground, talking to our clients, talking to our team members, talking to our partners, and making sure that we make the right pivots to deliver on the guidance of 2026 and the same on the long-term guidance.

With that, we'll turn it now to questions from the audience and we're ready. Thank you.

[Operator Instructions] The next question comes from Remi Grenu from Morgan Stanley.

I've got three on my side. The first one is on Core Services momentum. So I think you're flagging the geopolitical uncertainty. Can you give us a little bit more flavor on the level of activity in March versus the rest of the quarter. Just want to understand if there has been a deterioration since the beginning of the Middle East situation. And how you would expect Q2 to look like if we put together this macro uncertainty, the offshoring and the overall pipeline for contracts. So that's for Core Services.

The second question is on LLS. I think you're a little bit more positive in the press release saying that the momentum has improved through the quarter. Just want to understand if it is just March being better on the back of the easier comp base? Or are you seeing a daily pickup in volume of activity that would make you more confident on that part of the business?

And the third one is on the transformative deals and the outsourcing from clients. I think you were as well referring to a few of these deals back at the full year results presentation saying a few hundreds of millions of euros, potentially larger deals, transformative, et cetera. Can you just make an update on pipeline and if you expect these to come through at some point?

Excellent. Thank you, Remi. And maybe the answer to question 1 and question 3 are related in a way. Definitely, we continue to hear, and we continue in conversations with many of our clients for these large transformative deals. The ones that we knew at the last time we chatted are still there, and we continue talking to our clients as they go through their own internal changes, they go through some uncertainty in terms of deciding to move ahead given the geopolitical environment, particularly with rates going up and down from a financing perspective. So that has been a little bit correlated.

Just to finish on your first question on what are we expecting? Yes, we did see a slower ramp-up, particularly in EMEA that you see that in the results as well, we saw a slower ramp-up of some of the contracts towards February and March. And of course, that we hope that now in April, May and June, we see a pickup on that. But given the situation, it would be too adventurous for me to say with full certainty that, that is going to be the case. But that is something that we continue monitoring. And hopefully, some of the macroeconomic situations get better for all of us. And that at the same time, that impacts the business. So those are the 2 sides of the coin. So that's why I think your question 1 and 3 are interrelated.

On the question on LLS. Definitely, we see momentum. So March over March is better, but we have seen that momentum growing all throughout the quarter. So every month was a little bit better, not only than the previous year, it was a little bit better than the last month, and we continue to see activity picked up in that space. We've also had some wins of large hospital systems and other type of clients in the space of LLS. And that gives us the confidence that we see the strength building up on LLS in particular, but that is what we're seeing there.

The next question comes from Nicole Manion from UBS.

I also have three, please. Just the first one around your comments about monitoring the space on content moderation. Could you maybe sort of spell out a bit what you mean there? Are you sort of assessing your position in this part of the market?

Then the second question, just on offshoring. Obviously, you've been seeing very strong demand for offshore for the last few years now. But obviously, that means that base is moving higher all the time. Has the rate of growth in offshore sort of still been accelerating at this point? Or are you seeing kind of any plateauing or moderation at this stage?

And then finally, I appreciate this is a sales update, but can see you've reiterated the guide on restructuring of EUR 70 million to EUR 90 million, I think at the time of the full year update. So as of end of February, you talked about EUR 56 million in costs year-to-date that have sort of been committed. Just wanted to ask about why these costs are so front loaded? And if there's any sort of risk to the upside on that restructuring amount given the ongoing change around AI, obviously, a new Chief AI officer in the group and so on. Any detail there would be really helpful.

Great. Thank you, Nicole. So let me start with content moderation. Look, definitely, we're seeing the continuation of the trend of automation in this space. We're also seeing and monitoring very closely some of the developments in the U.S. with regards to content moderation and the role that social networks are having in lives and livelihoods in that space. So that is the space that we're constantly monitoring.

In my conversations with some of our clients in this space, they are telling us that, of course, as a result of some of the developments that we've seen in the United States, they are getting increased regulatory pressure. And as such, how do they balance their human moderation vis-a-vis their AI moderation? And that is a space that we continue to have a conversation with them in ways that are not only, hey, of course, there's automation that can go and continue in that space. But at the same time, how do you get the human in the loop in that space, given some of the developments that we have seen, given the effects that content moderation can have, what is the impact for their own operations.

So we're very, very active in these discussions with our clients. And again, we are seeing a little bit more of the pressure for them in this space. So definitely, that is what we're monitoring. And of course, we will continue to assess that space. The second question was on?

Offshoring is accelerating or not.

We are seeing a higher rate of offshoring in Q1 in that space, probably a little bit higher than what it was before. The rates -- we continue to see very healthy growth for our India operation, as Benoit said, South Africa, Egypt, so I'm not seeing particularly a big pressure on the pricing side in that space, but we continue to see very healthy growth in some of those operations for us.

And then the last one was on the restructuring costs. Yes, I think, Benoit, you can confirm with me. It was EUR 56 million.

Yes. EUR 56 million was corresponding to what was announced at the time we released our numbers. And your question was why is it front-loaded? The answer is it's better to restructure and get the savings on a full year basis. So effectively, the plan has been set so as to invest as much as we can early in the year and so that we expect to see the benefits on our margin by year-end, and this was factored in our EBITDA projection for 2026.

Exactly. So that -- hopefully, it answers your point on the restructuring. And as a reminder, also, all our -- most of our AI expenses also flow through our P&L as OpEx. So in terms of any CapEx or restructuring costs that we need to do that, we can manage it within our OpEx guidance.

The next question comes from Suhasini Varanasi from Goldman Sachs.

Just a few from me, please. You've done some reclassification of activities. Can you remind us when TP Infinity was recognized previously? And will you be giving us the margin reclassification? It will be helpful to have the quarterly revenue numbers, but maybe just the margin data as well, that would be helpful.

Second question is you've had some time to think about business. As you think about transforming that, do you see scope to unlock value through maybe some divestments, disposals, et cetera? I would love to get some color there.

And then I think the last one just to get some clarity on the guidance again. I think you indicated that you still expect a soft first half, an improvement in second half of the year. So does this mean that we should expect declines overall in first half?

It was a little bit hard to hear you at the beginning. So let me make sure that we got the question right. And I think the question was if we're going to give also the margin breakdown for some of the reclassification that we completed in Q1. Is that correct?

Correct. And just the rationale behind it, please.

Yes. No. So on the rationale for the reclassification of TP Infinity, like I said, on TP Infinity was one single unit of our core service reported mostly in EMEA because 90% of the revenues were from EMEA and managed in EMEA. We expanded our operations in the U.S. starting last year. You have the breakdown. You see that it's around EUR 3 million per quarter. You have the breakdown of revenues in the press release. And because this is growing, we expect it to effectively be reported separately and it reports under America with core services in America now starting in January of this year. So this is why we align it.

I don't think we will disclose separately the margin on this business because, for us, it's one of the component of the business. So it will still be reported as part of core services for the moment unless that becomes a very material part of the business and with a different dynamic, but it's not the case yet.

Exactly. I think your second question was a quick update on the portfolio. It's a strategic portfolio review. So with that, what I can tell you is that, that exercise if you remember, with some of you that I've discussed a few weeks ago, we have the mandate from the Board to conduct that strategic portfolio review. The portfolio review is in full swing, and we are already deep into the analysis that, that has. And before someone else asks me, I will put it out there. Of course, some of the decisions that we make in that space will have implications on some of the capital allocation decisions.

So we are looking into that space all as a block. And I will come back to the investor community when I have more news to report in that space, but it is ongoing. We are looking -- we have discussed with a number of partners, and we are having very, very interesting conversations of different ways to unlock value, but nothing firm that I can report as of right now. So I will come back to you on that one.

And I think I'm missing one question, right?

Was the soft H1 and whether we expect a growth in H1 compared to the overall guidance.

So we are expecting a softer H1 as we said, and we're working day in, day out with our teams and definitely an acceleration of our growth to hit our guidance in the second half. So with that, you can make the estimates of where we are going to be in the second quarter of the year with a softer start and then the recovery on the -- the upswing, not the recovery, the upswing on the second half.

So just to clarify, so you don't expect the inflection of growth to happen in the second quarter. It's more a second half weighting.

Yes. We see an acceleration of the momentum of all the actions that we're putting in place. But of course, we expect them to materialize as expected on the second half.

[Operator Instructions] The next question comes from Karl Green from RBC Capital Markets.

The first question is just about the outsourcing dynamics, which I think you said in your narrative, clients are regarding as a core strategic lever. And just in terms of that particular point, could you drill down a little bit and explain what they're thinking at the moment, whether it is a core strategic lever because they're thinking more about the offshoring or it is that technology integration? And linked to that, have you had many instances of clients who have gone away and tried to implement and integrate AI solutions themselves. And it's not worked out quite as well as they would have hoped and then come to you for support and realize that they can't go it alone. That's probably two parts to one question there.

And then just the second question, much more straightforwardly. Just in terms of the trust and safety business, just how far off peak are the activities in both the Americas and Europe. I mean it seems that Europe isn't quite as far down the automation process at the Americas. But just to get a sense as to basically how much further either of those businesses have potentially got to fall, please?

Yes. So let me start with the partner and the strategic discussions that we're seeing there. The clients are coming to us, and they are saying, like, hey, definitely AI is going to reshape some of the processes that we're doing, some of the operations that we have. We certainly cannot do that with a multitude of partners. We want to do that with a few strategic partners that get us, like that's were a quote that I heard from one client that's like, you guys get us. You guys have been with us for so many years. You know our processes. You know our data, you know the kind of things that we need to do and the kind of things that we don't want to do with our clients. So we want to partner with you. And we want to partner in these deals where they see outsourcing as a strategic lever because in many cases, we are one of the primary partners for them. In other cases, they see the strength of our offers vis-a-vis some of our competitors in this space that we're able to offer both the reliability, the security, the global reach, the set of AI solutions that we have.

So they are coming to us and saying, like, hey, it would be too expensive for me to do it internally or it would be too complicated for me to do some of these things internally. I want to do that with a reduced number of partners that are more strategic and that are there for us. So that's a dynamic that we are seeing. As part of those conversations we've had and I've had a few of the partners, our clients telling us, okay, we tried these AI solution. We tried this knowledge management system. We tried this Agentic voicebot. And we really didn't get the uplift or the reduction that we were hoping. We would really benefit from having your expertise in our processes, your expertise into what you do and help us drive a little bit more.

And again, a little bit of the conversation in that space is not only on, let's do the same with fewer humans and this is the part that excites me the most. It's like, hey, help me do things that I cannot do today. Help me make sure that I reach out proactively to clients when I know I'm going to have an issue. Help me do collections in a very different way than what I'm doing today. So these are the kind of conversations that we are having. So it is not only about like, hey, help me implement the tool. It's helping me unlock the value that comes with the tool. Of course, there's a technology component, but of course, there's a process expertise. Of course, there is a knowledge of the strategy and the objectives that organization wants to drive.

And that's the most exciting part, particularly in a world, and I'll emphasize it again, in a world where we are seeing that the technology is becoming much more accessible and that a lot of the technology that is out there in the market is more and more of a commodity. And therefore, the true differentiation comes from the people that know their business versus the people that just have a technology capability. So that is the space and the conversations that we're having.

If I go to your other question on trust and safety, the reason we are seeing it more in Americas for just a simple reason. I think a number of the LLMs are better trained in English than in some of the other languages where we perform some of the content moderation and trust and safety activities. And therefore, we see more impact first there. It doesn't mean that it will not get to some of the other languages, but that is what we are seeing. And again, it's a space that we see already the evolution given some of the developments in the U.S. in that space and the increased regulatory pressure that some of our clients are getting.

So we will continue to monitor that. We will continue to make the investments on getting into the real high value-added activities without -- within trust and safety, but that is a little bit of what explains the delta between Americas and EMEA and APAC.

The next question comes from Ben Wild from Deutsche Bank.

Two questions, please. The first one is hopefully a simple one. Can you help us understand roughly what the unit price differentials are between onshore voice, offshore voice and AI solutions within Core Services today?

And then a second question just on the free cash flow. There's a comment in the press release that FCF will be H2 weighted this year. We've already discussed the H2 weighting to growth and also the H1 weighting to nonrecurring restructuring cost items, but are there any other specific items that mean FCF will be H2 weighted and what kind of split between H1 FCF and H2 FCF would you expect?

Yes. The price differential, I think this you can also get from multiple sources in terms of what we see in the delta on price, 20%, 30% or even depending on the geography, up to 50% difference from a unit of a cost per minute if we can still use that metric in an AI-enabled world. But that is just to give you a rough sense of the onshore versus offshore.

The AI solutions, look, we are partnering with many companies, and each of them has a very different way of pricing it. Some of them are pricing also per minute. We are seeing some of those prices per minute be at parity or higher than some of our prices at offshore locations, mostly and even in some onshore location is getting close. But we are seeing that. So we see the price per minute. We see the price per token or consumption. And there are a very few selected number of partners that we're really excited about that they are partnering with us on per resolution or per outcome basis. So it's hard to say it's x percent cheaper or not because the basis for comparison on that are different given the way that they are pricing it.

Again, in the places where we see a price per minute, we see that is getting closer and closer to some of the offshore prices that we can get. And if on top of that, you get some of the very interesting developments that we are seeing right now on translation for certain parts of the business, then it starts to become a more complicated equation to solve in that space. But that is just to give you a sense a little bit of the dynamics that we're seeing from a pricing perspective. Benoit, do you want to comment on the restructuring costs?

Yes. But normally, we don't comment on that for Q1. But effectively, we have incurred big part of the restructuring costs in Q1. We will continue to do so in the coming months. When it comes to the balance of the free cash flow, as we mentioned in the press release, we expect to have broadly the same type of allocation that we have seen last year. I mean, net of those restructuring costs, which is a slow free cash flow in H1 and most of it coming in H2 with the same type of seasonality and roughly the same percentages.

There are no further questions at this time. So I hand the conference back to Mr. Amar for any closing remarks.

Excellent. Well, thank you for the questions. I was expecting them to be very vigorous as always, and I'm sure you will have a few more over the next few days, and we remain at your disposal to answer any additional clarifying questions.

Again, we're reporting a Q1 that is first, as anticipated; and second, not where we want to be. But I want you all to have the confidence that we are working very, very hard day in, day out since I took over as CEO to get us on the right track, to accelerate the transformation and to be able to report and I really look forward to coming to these calls to report some of the progress that we are seeing that we are already working on that we're executing throughout the company from a revenue perspective, from an AI perspective and now welcoming Andreas to the group and also from a cost and profitability perspective. So more to come.

I look forward to sharing some of those updates with you. And with that, we will see you at the end of the first half in our next call. Thank you very much.

Welcome to TP 2025 Annual Results Conference Call. [Operator Instructions].

Now I will hand the conference over to Thomas Mackenbrock, Deputy CEO. Please go ahead.

Good evening, everybody, and welcome to our 2025 results presentation. And as you have probably seen, we have a lot of news to share. As always, with me in the room is Olivier Rigaudy, my dear colleague and CFO of the group. And we have a special guest today, Jorge Amar, our incoming new CEO for the group who will present and introduce himself later today.

But let's first have a look at the agenda for today's call and what we will cover in the next 50, 60 minutes or so. First, I will give you an update on the key highlights of 2025, provide a strategy update of where we stand today with the implementation of our future forward plan and an outlook for the future. Olivier, as always, will cover in detail the financial results. And at the end, we have ample space for Q&A.

Let's look at the key highlights, and let's focus first on the financial aspect and then talk about in a second step about some of the strategy and governance changes we're seeing. 2025 has been a turbulent year for the world and for our industry, but we as TP have delivered solid results. And as you have seen in our press release, we have met all our updated 2025 objectives.

If you see on the group revenue, we are reporting again a bit over EUR 10 billion in net revenue, and we have grown on a like-for-like basis, excluding the hyperinflation effect of 1.3%. If you exclude that 1% on a reported basis, given the weak U.S. dollar minus 0.7%. It's particularly noteworthy, and we talked about this in our Q3, our H1 and our Q1 presentation that our Core Services are a stable growth momentum and a stable growth anchor for the group, with reported 2.7% like-for-like growth, which is remarkable in this environment, while at the same time, our Specialized Services division faced some unique challenges last year.

On the profitability, again, we delivered our updated 2025 guidance. We have reported an EBITDA of almost EUR 1.5 billion, with a margin of 14.8%, excluding the currency effect, which means on a reported basis, 14.6% and this also translates into a very healthy net free cash flow. If you exclude the nonrecurrence of over EUR 900 million, and we had a record cash flow generation in the second half of the year with more than EUR 640 million. So we are quite proud about the results in 2025.

And when we look at 2026, we provide the following guidance. For this year, we expect a growth rate again between 0% and 2%. But given how we started into the year and how we ended last year and given some of the uncertainty, in particular, in the core onshore market, the U.S. and Continental Europe, we anticipate for Q1 a revenue development, which will below the annual guidance.

Secondly, and you will see later in detail some of the measures we are implementing. We also expect a stable EBITDA margin, which means 14.6% on a reported basis that assumes a dollar of $1.20. The net free cash flow generation is expected to be this year slightly below last year, given the strong euro. And so we see here a range between EUR 800 million and EUR 850 million excluding the nonrecurring items. In our proposal, we just had the Board meeting this afternoon to the annual shareholders assembly at the end of May is to increase the dividend from EUR 4.20 to EUR 4.50 per share. That's on the financial side.

Let's take a look at some of the governance and strategy updates. So we are very happy sort of to announce the long-awaited process of our governance change. The Chairman Moulay Hafid, Daniel the founder and CEO and myself has recommended to the Board and the Board sort of followed that recommendation to appoint Jorge Amar, who is a very world-renowned AI expert and leader of McKinsey's global customer service practice, to be the new CEO of the group. He will start officially March 16. I've known Jorge for quite some time, and I'm very excited that he steps into this role.

This also means naturally that Daniel, myself and Olivier will step down a day before. Daniel will also step down from the Board of Directors. And we also, at the same moment to really make sort of the governance renewal complete also co-opting to the Board for new members. One will be Jorge starting middle of March, sort of stepping into the role of Daniel. Myself, also, I will continue to support the group that is very close to my heart but then in a different role as a Board member and 2 very exciting new Board members who have been co-opted and are then up for the approval by the shareholders' assembly at the end of May, one lady from Qatar and one lady from South Africa, which I'll explain later her qualifications.

So that's really, I think, quite exciting news. I have been many discussions over the last years, when will this happen? I do believe we have there the right team on the start, and I'm excited sort of to support this group in this new role and particularly Jorge in his new task.

Then Future Forward. We launched this initiative last summer, you saw in Q3 a quick update. We are now in full swing and 2026 will be the first full year of implementation. We have really mobilized the organization with hundreds of different initiatives. And as I explained and hinted towards in our Q3 presentation, we are working strong on essentially 3 levers. We want to accelerate the growth. We want to drive efficiency also leveraging AI, and we want to transform the company and we are making sort of good steps on all 3 elements.

And on the internal AI efficiency, we are starting a program as we speak that will drive efficiency savings for the group targeted to be over EUR 100 million in 2026. We have launched more than 500 AI projects last year and expected to scale further with our TP.ai strategy, and we are also happy to share that also under the new governance, we are launching a comprehensive strategic portfolio review of the group. So a lot to be discussed.

Let's quickly look at the highlight numbers. I think no surprises here for the audience, of course, happy to answer more questions, but we see the strong Core Services that we saw throughout the year. There has been a little bit of weaker momentum in Q4 that we anticipated in our November presentation what for me particular positive is to see the momentum in the Americas. As you remember, it was a bit negative before, but we have seen an excellent development in India as well as Latin America. And given the strong momentum, we are reporting growth of 1.4% like-for-like in the Americas.

In EMEA, again, very strong with close to 4% in 2025. We have seen great momentum in the U.K., South Africa, Egypt, APAC as well, sub-Saharan countries. So they're across the Board a very strong momentum, while also a bit subdued in Q4.

Specialized Services, on the other hand, you know the challenges on the nonrenewable of the significant Visa contract and the market environment for our Specialized Services in the U.S. So from that perspective, minus -- a bit more than minus 9% like-for-like growth. If you adjust for the effect of the Visa services contract, we have as indicated and as expected, a slight positive like-for-like growth for Specialized Services. But important to note, yes, the momentum has been reduced but given all the measures we have taken last year, we have proven to maintain a strong profitability. There's only a slight decrease of this highly attractive business.

Second comment, again, in times of uncertainty, having a broad client portfolio is key and giving our broad exposure to multiple different industries has been and will be a strength of TP. For 2025, we continue to see strong momentum in public sector, fast-moving consumer goods and strategically very important, the strong sector of financial services and insurance. This is really has been sort of supporting the growth last year, we saw a bit of lower activity, automotive and energy utilities last year.

Also, the portfolio we talked about a lot, TP is not a company that stands still. Over the years, always have been able to develop new business lines and build out -- building on its capabilities, new services line, along with our articulated future forward plan. And you remember the presentations last year, we have seen strong growth momentum in AI data services, and we call this out for the first time. We have seen very strong high single-digit growth in sales, which is 7% of the group and which is a critical factor to provide Revenue as a Service for our clients. And also very strong momentum, double digit actually in our back office and BPO-related task, which is important to sort of have an end-to-end service chain for our clients.

Trust and safety. As indicated, we saw some revenue decrease. There is some automation happen on our client side. And care overall, broadly in line with the overall Core Services growth, so also a healthy development but changing the way we operate for our clients.

Now a quick update to our 2 new executive managers. Jorge Amar, as I said, very happy, I got to know him very closely really now for quite some time. He has been working with the group for quite some time. But Jorge, why don't you introduce yourself to the audience and to our investors.

Thank you very much, Thomas, and thank you for the warm welcome into the group. Today is not the day to speak at length as I will officially become the group CEO starting March 16. But as a quick introduction, Jorge Amar, I was born in Argentina, but most of my professional career has been in the U.S. where I worked with some of the largest companies in topics around customer experience and service operations. And in particular, over the last few years on the topic of artificial intelligence. Not only from a technology perspective, but also how to think about consumer and employee adoption. So all these feels like the right combination of things that are leading me now to be very proud in joining the group. So again, more to come starting March 16, but very, very excited to join the group.

So I think not just Jorge is excited, the entire group is excited. I think it will be a great addition for the company. And as he indicated, he brings 3 key components that are critical for the group; first, a deep understanding how AI works in enterprise environments which is absolutely critical for our journey ahead; secondly, he has a strong proximity to existing and potential clients of TP, understanding their needs, understanding their environment, having these relationships, which I think is super critical also for our path in the future; and thirdly, obviously, given his background, he has a strong analytical mind and sort of will shape the strategic path for TP in the years ahead.

I can also -- he's not with us today, but also can only praise our new interim CFO, Benoît Gabelle, has been a Deputy CFO for TP for some years now. Before he was advising the group, he was a partner at EY, is an absolutely excellent person. We are very excited that he will sort of step up into this new role and will support Jorge from the financial side.

Also, as I said, it's not just the executive management team, but also the Board has renewed and has co-opted today for new members, 3 of them immediately. Sheikha Hanadi bint Nasser Al Thani, a very renowned Qatari entrepreneur, investor and business leader. We are very excited that she brings her expertise, her network into the board realm for TP. She has strong expertise when it comes to investment and the investment in capital markets.

Secondly, Ingrid Johnson, she's South African lady, also with a broad understanding about capital markets investment but also the banking insurance space where she led several companies. So quite excited for that sort of additional expertise on the Board as well.

Jorge Amar will join middle of March. And as I said, I'm very dedicated to the group, and I'm excited to continue the journey with the group in this new role as well. Of course, all of these cooptations are subject to the shareholders' approval at the meeting at the end of May.

Now let's look at the numbers, and Olivier will guide us through.

Thank you, Thomas. Good evening, everyone. I'm happy to present to you the 2025 figure. As mentioned by Thomas, I do believe that -- we do believe that we have delivered a very good year despite this global challenging business environment. As you can see here, you have the full P&L. But before commenting in detail, I just wanted to highlight 3 topics.

The first one that was unexpected when the year started. The macro environment has been difficult all along the year and the growth at different market was probably lower than we expected. Secondly, we have the FX environment that was not really exactly what was supposed to be to happen. For you -- I remember that we start the year with a dollar that was at EUR 1.03 and finish it at EUR 1.17. So it has been a global wash all along the year, especially in the H2, we will come back in a minute to that. That was not exactly the plan. And lastly, the impact of the Trump administration policy on our major business of Specialized Services, I was thinking, of course, of LLS has also an expected impact on the growth that we were supposed to deliver this year.

But beyond that, beyond -- despite that, we have been able to post a sales figure of EUR 10.2 billion, 1.3% like-for-like growth, excluding impair inflation. And EBITDA, which is above EUR 2 billion and EBIT before nonrecurring items close to EUR 1.5 billion, EUR 1.485 billion aiming to 14.6% growth rate -- sorry, to sales versus 15% last year. I'll come back to explain where I come from the difference.

Finally, the net profit -- the operating profit is roughly equal to last year. We will see why. We have been able to reduce tax charge significantly, and our net profit is roughly the same than last year. As you can see, this is a 40 basis point difference in EBITDA margin, which -- of which 20% -- half of it is coming from the FX.

Let's have a look to the figure of sales first. The first thing to tell is that, of course, when you start to have a look to the figure of last year, you start with 10 -- also EUR 10.3 billion and you have a currency effect of EUR 362 million, of which EUR 240 million came in the second part of the year. So you had a 50% increase of the negative impact in the second part of the year that was significant.

So when you start to look at the precise figure the way they have been built, you have also, of course, a change in scope of consolidation which is a consolidation of ZP better together. You remember that we bought this company last year, and we consolidated early February 2025. And we are also a small company called Agents Only that came on board early July 2025. So you have a positive impact of scope of EUR 196 million covering the decrease of Specialized Service, EUR 132 million that was mentioned by Thomas a minute ago, of which most of it is coming from this U.K. contracts that we have not been able to renew last year. That has a big impact on our sales, EUR 140 million to be precise.

And beyond that, the Core Service business -- the Core Service activity have been able to grow by close to 3%, 2.7%, which I believe is beyond the market figures that we will get in some weeks from now, showing that this group has been able to continue to deliver significant growth in different markets. It was mentioned by Thomas in U.K., in different sector, in public sector, in banking where we are able to match the demand of the client.

Let's have a look to what happened specifically in this year. When you look the FX environment, things are clear. You have all our currency in which the group operates have been degraded versus last year. So it has an impact. Of course, the dollar, but not only the dollar, the Indian rupee, the Philippine pesos, Sterling, everywhere. So we are facing a situation where we have been able -- we have not been able to cover, of course, all the translation effect that has a final impact on our mix of margin. This is an adverse FX environment in 2025. That was effectively significantly higher than people who are waiting.

If we look now to the result by, I would say, by sector or by zone and by activity. I would say -- I would -- I'll add 2 points. The first one is a strong EBITDA margin improvement that we have been able to do in Specialized Services in H2. You remember that in Q1, specifically, but also in H1, LLS has been hit by this Trump effect, if I may say. But the group has been able to react quickly and to adjust this cost quickly to match the global demand. So the demand is flat versus the volume is flat in Specialized Services, notably in LLS but we have been able to recover significantly the margin, and we have just a small negative effect for the full year that is going to be positive next year, again with LLS given the measures that has been taken all along the year 2025.

When it comes to Core Service, there are 2 issues to be -- to have in mind. Of course, the FX impact, which is -- I just mentioned it very significant. And the group decided to put some money, some investment in AI and IT technology that has been, I would say, spend notably in holding, as we can see on this table to support the future growth that was absolutely needed for the future.

So all in all, the result in margin are not dramatic if you look at it. They are much more -- they're positive. If you look what happened, you have versus last year, an impact of Specialized Services that is roughly neutral. Of course, we have lost 70 basis points with the TLScontact impact, notably the UKVI -- the U.K. contract, sorry. That has been covered by 2 things. One is the acquisition of ZP that came on time, and that has been made on time accordingly and also by the mix effect linked to the work that has been done all along the year with LLS to improve the margin. So all of that, meaning that the cost on the Specialized Services impact on the margin is neutral and has been -- and we have been able to swallow all the impact of the TLScontact that we lost.

Beyond that, you have the 20 basis points that are linked to the FX roughly. And you have the 15 basis points, which are the costs related to AI, notably spend in holding, as I mentioned earlier on. So I do believe this delivery of EBITDA margin is really good and shows how the group has been able to adapt to this global environment, either in terms of demand for LLS or either in terms of adverse FX condition across the board.

If we now move to the other part of the result, what we can say is that the amortization of intangible assets are flat versus last year and the nonrecurring items are a little bit better than last year. You remember that last year, we had a significant amount of money that was spent to deliver the synergy from Majorel. Of course, this year is significantly less. But we have been able to -- we have been obliged to get out of some country, of course, Russia, that was one of the actions that we did all along the year, but also 2 other countries like Guyana and Trinidad, where we wanted to get out.

Besides that, we have been careful on the impairment of some assets, notably on PSG which is recruiting activity that we bought 4 years ago. And where we are really, I would say, cautious on the future market for 2026. And we thought it was clear, better to be cautious and to impair at least EUR 60 million -- EUR 67 million for this business. It doesn't that mean that the business is not good, but we are very, very careful here. I remind you that this impairment of goodwill has, of course, no impact on cash.

So the operating profit is roughly flat, EUR 1.55 billion versus EUR 1.82 billion last year. And when you look what's happening on the final part of the P&L, we have been able to maintain our net financial charge at the same level despite the fact that we have an outstanding debt that was increased in the year. But of course, last year, you remember, we had a very, very positive hedge impact coming from the devaluation of the Egyptian pound that didn't happen again this year. The impact of this hedge was EUR 50 million that is not happening again. So besides that, we are flat in finance cost.

What is interesting is that we have now finished -- mostly finished the integration of Majorel, and we have been able to reduce significantly the tax rate -- the accounting tax rate. The impact is EUR 56 million improvement in 2025 versus '24. And we are still more things to come and the full year effect of the decisions that we took and implement in 2024 and 2026. That's the reason why we believe that in 2026, our tax rate will be below 30%.

Beyond that, very few things to tell that we are roughly at EUR 500 million at net profit level versus EUR 523 million last year. Remember, we impaired EUR 67 million from PSG, that has a big impact on the net profit.

More interestingly, and it as was mentioned by Thomas a minute ago, is a strong free cash flow generation. You remember that was a question about our ability to deliver free cash flow for the full year following the performance of H1 that was hit by some one-offs that were, I would say, exceptional. We have been able to deliver the best cash flows that we ever had in the H2 -- in the second part of the year in 2025, EUR 642 million versus EUR 636 million for the following year -- for the previous year, sorry.

We did that because we manage strongly the working capital management or strongly working capital as expected. But we did that without cutting in the CapEx. And that is absolutely key. We continue to invest reasonably, but clearly, in some place where the demand is rising, notably India, South Africa, where the market is asking for size and for volume. So we increased our CapEx to 2.4% sales -- to the sales this year.

So at the end of the day, the free cash flow is at EUR 900 million, EUR 901 million. Keeping in mind that we have to pay, of course, remember that we have the French restructuring plan, voluntary restructuring plan that was partially paid in 2025 for EUR 25 million out of the EUR 31 million that are shown here. And of course, will continue to be paid in 2026. So as a whole, strong free cash flow generation, I know it was a concern about the market, but the company continued to deliver strong free cash flow, and will continue to deliver strong free cash flow.

If we now move to the situation of the group in terms of balance sheet. As you can see, we have been able to stabilize the debt roughly at 2x -- below 2x net debt to EBITDA while returning to the shareholder 42% of the free cash flow through dividend and share buyback and continuing to invest in business. I just mentioned it a minute ago, but also acquiring ZP and establishing some AI partnerships that are going to be promised -- promise for the future.

So all in all, we continue to have a strong balance sheet while continuing to develop the business. And when you look at the indebtness, there is no reason to be afraid. We are BBB rating -- Standards -- S&P. We are the -- we have launched -- I remember you that we launched early last year, a bond of EUR 500 million that has been easily covered by the market. And we have a debt that is, I would say, balanced between the financing source and by nature of rate.

To be clear, the group has the ability to reach -- to have access to lately between EUR 3 billion and EUR 4 billion easily through commercial paper, through some medium-term bond or banking facility. So the average cost of the debt is below 4%. We have an average maturity, which is around 3 years and we are absolutely confident about the ability to continue to finance and support the business and the growth of the business in the future.

That's what I wanted to tell you. I'm holding back to Thomas for the strategic part.

Thanks, Olivier, and thank you also because this will be your last presentation to present in your results after 16 years with the company. So a big thank you on behalf, I think, of the entire Board, the entire organization for these wonderful sort of decisive action over the last 16 years.

Thank you.

Let's look, and -- I'm in the interest of time, quickly as an update on Future Forward that you see where we stand and what will be continued. So as I said, the value creation office for Future Forward is in place, hundreds of initiatives activating. I brought for today's presentation, as promised last time, 4 examples, to give you a little bit of a flavor where do we stand and what is happening and to have a little bit more tangible view on these growth levers, transformation levers as well as efficiency levers.

Internal AI, we talked about, we see 3 big levers on driving change in the organization, of course, leveraging AI in everything we do internally when it comes to recruiting, training, workforce management, supervisor quality, but all corporate function, if you will, and AI adoption allows us to reach another level of quality, but also efficiency. Hand-in-hand with this internal AI transformation goes the cost optimization addressing structural changes through delayering automation on our SG&A and our overhead parts as well as on our direct costs as well. There are many, many plans in place now that are being implemented and they allow us to drive the savings that you see below.

And thirdly, that is part obviously of the new leadership role with Jorge to find a simplified organizational redesign and to choose some lever there to have a more agile, leaner organization. Overall, for all of these 3 levers, the current expectation is that this will be delivered above EUR 100 million run rate savings, and we expect a onetime cost this year, of course, on depending negotiation of some of the levers between EUR 70 million and EUR 90 million. These plans are already in action.

If you look at our annual results, you see that in January, February, we have the first measures amount with a corresponding cost of EUR 56 million. So it is happening. It's being implemented, and it will be continued seamlessly also by Jorge in the future. So this is on track and in execution.

Second one, transformation. All of you remember this chart what I presented in Q3 that we as TP, believe AI is not a piece of software that is being sold. It is an incremental part of our operating fabric to drive outcomes for our clients. This is true on the functional side, so industry agnostic, and we have made good progress on some of our functional solutions, as you see later, as well as of the industry solution side. You need to orchestrate like we do today with TOPS and BEST, the human dimension, you need to orchestrate the AI dimension as well that it really can unfold this ROI and impact for our enterprise clients because otherwise, it's just a nice demo, but not really something delivering value.

For this, we have started, as you remember, at our Capital Markets Day, our Q3 presentation TPI fab, our foundational backbone, we've launched more than 500 AI projects this year, integrating what we have done in the past into our new solutions suite in really driving impact for our clients. The biggest impact because there we had a head start in the past is augmenting with AI, our existing human delivery engine. There, we have seen more than 270 projects last year of doing this human augmentation but we also started to see some traction on FAB Connect, which is basically orchestrating human and agentic AI; FAB Growth, enabling with AI revenue as a service for our client; and FAB Collect, agentic AI collection where we see a lot of potential. This is a journey that will basically carry on the next years ahead but the foundation is laid, we are continued to developing.

And the examples are real. Wherever you look, whatever new proposal you have, whatever new win you have for a client, AI is part of our offering is attached and ingrained what we do today, whether this is for a leading health care insurance company in the U.S. where we build an AI-based tool that allows faster access to the knowledge base. We won a client last year in Asia, it's a large bank, where we integrated human customer support with agentic AI customers on board to manage high-volume cases. And at the same times, this orchestration between human AI and agentic AI was the winning case that the client entrusted their most treasured valuable resource, their clients to us with our FAB Connect solution.

We have won a large telco company in Latin America, where we do agentic AI collection. So we can be earlier on in the building cycle, reach out with an agentic collection tool and then hand over in complex cases to a human. And this is an example, again, where is the value add for TP. We are knowing which AI technology is available in the market, depending on the situation, depending on the client need to plug it in our processes. But as we work with dozens of different telcos in different countries, we work on many different debt collection services. We have the data now how do you orchestrate the process to unfold the power of the AI. And we've seen great results after the implementation actually quite recently when I visited the client.

And lastly, FAB Growth. 7% of our business today is sales. There, we are not a cost center, but a revenue engine for our clients, and it's obvious, but it's hard to implement how AI augments our humans to drive better sales for our clients. We have started working for many high-tech companies in that felt with really incredible success. And I see there's really a great momentum combining the human power of sales with the tools of AI.

Maybe in the interest of time, just a quick sneak preview, and I'm sure you will see in the next years more from Jorge and the team. I really believe if you think about and sort of cut through all the noise in AI, finding the right recipe, how you orchestrate in a world where AI is ubiquitous, the human power with the AI power is key. It's not just about load balancing. This call is done by AI, this by a human. It's about understanding where hallucination happen, how do you design the data flow, where does AI play a role for better outcomes and maybe a human how do you manage this handover.

We're investing quite a lot right now of building this tool, including in a responsible control center that can detect hallucination, accuracy problems, false answers, defines the right guardrails and really configures outcomes for the client. TP is not a company that is selling AI solution. We are a company that drives outcomes for our clients and managing the orchestration of an operating machine. And the operating machine has a human hand and an AI hand or AI leg and doing this orchestration in the right way is key in the future because our clients don't want to see a demo or buy a tool like in a software, they want to see an enterprise process managed with a measurable impact. And that's, I think, the role for TP, you will see more in the future, but it's on the move. It's being developed, it's being deployed in client places, and I think we're all around the table are quite excited about it.

Then many of you asked what is happening? How -- can you show us more concrete examples for sales? I talked about it, 7% of the group, EUR 700 million in sales. We do B2B2C and B2B2B sales. Starting with high-tech clients. We invested last year and the team built it out to not just focus on high-tech block, fast-moving consumer goods, banking, telco with really some good traction. We've seen high single-digit growth last year. We expect nothing less this year from the team, and you see it's again, this blend of human talent with AI.

And the same is true with data services for AI. We called it out now it's 2% of the group. I think we all wish it will be a higher number but we see double-digit growth with the team. It's a market that is growing. It has moved from general data labeling and notation based on general knowledge to way more specific needs, way more specific expertise for clients, really combining domain expertise on certain subject area experts and bring it again for enterprises to life and having enterprise solution for medical companies, for car companies, for banking companies and combining on our know-how is quite critical. We won their 5 new clients. And again, the expectation for this year is at least to continue the growth momentum we've seen in 2025. And with this, I think these 2 examples, it shows you how the portfolio of TP is changing over time.

Last but not least, outlook. As you all know, the world is uncertain. Our market is uncertain. If you look at last year numbers, we expect a growth more or less in the same range, 0% to 2%. Based on how the year ended and started into the year, we expect Q1 to be a bit softer and to be below that guidance range. EBITDA margin with all the measures remained stable at 14.6%. Of course, assuming no major fluctuation on the FX side.

Cash flow, again, EUR 800 million to EUR 850 million, excluding the nonrecurring cash-outs. This is due to if you look at this year's numbers, which is a bit higher, due to the stronger euro versus the dollar and dollar correlated currencies because if you think about India, Philippines, LatAm, the U.S., of course, where cash is generated and translated to euro, the amounts might be lower given the current FX environment and the AI efficiency program that I talked before.

Overall, I would say TP is in a position of strength. We'll remain in a position of strength but needs to transform. Olivier, myself and I know also, Daniel, are quite excited about the future. We're stepping down, knowing the company in good sense with Jorge and are looking forward to any questions from the group.

I think you have seen this. This is the proposal for the dividend. Of course, for our investors is important it's being up for approval. May 21 in the general assembly. It's an increase of 7%, if I remember well, to EUR 4.50 the share, which is an increase and in line, obviously, obviously, with the position of TP wherein -- and the midterm guidance, there's no change there.

With this -- sorry, for that, open for Q&A, and I'm sure there are many.

[Operator Instructions] The next question comes from Suhasini Varanasi from Goldman Sachs.

First of all, a lot of changes, just trying to make my way through all of that. But maybe 3 questions, just to keep it short. When we think about the guidance for 2026, especially on the top line, can you help us understand your assumptions in Core Services and Specialized Services here and the implications that you're seeing on margins as well?

The second question is on the strategic portfolio review that you have announced. I see that you've taken a few impairments below the line in the last couple of years. Is that mainly in Specialized Services that you are directing with portfolio review? Or does it also encompass Core Services?

And it's interesting to see some of the color that you have talked about on Fab deployment. And it's good to see the benefits as well. Is it possible to help us understand the impact on contracted revenues and profits, margins, et cetera, as a result of deploying all of these AI solutions?

Okay. Let me start and then I hand over to Olivier for some of the impairment and financial topics. First one on FAB AI. If you look at the markets, Suhasini, I can -- I think it's too early to say what is the impact for the group. We are there in the beginning. It's part of the solutioning more and more. The question of, of course, how do you price some of these AI solution, how do you price some of the benefits. As we move forward, as we said in the past, there are some ideas to make this more tangible, but it's too early to tell what is the impact because we are also investing in the solution at the same time in terms of margin or not and in terms of pricing model in the future.

But you see there is traction, there's interest from the client. Every new offer that we have has a Fab solution inside. And let's say, I would be positive to see in the next 9 months, some more traction granularity that provides you also some facts that you can put in the model what the impact might be.

Strategic portfolio review, as also discussed in the past, of course, there's always the question on certain Specialized Services assets, but there is a clean sheet. Jorge Amar has the mandate for the Board to review the entire portfolio of the group. To be very clear, and as we put in the press release, including divestitures as well as including M&A. So both options are open. As I said here, he has a very strategic mind. I think many of our analysts has looked at the group, and he has a blank sheet from the Board also today to do a thorough review on the portfolio of the group.

Guidance, 0% to 2%. What was the question? We see a weakness -- we don't -- as you know, we don't give a guidance for Specialized Services and Core. I think the story of 2 tails that we have seen in the past that the Core shows higher growth momentum than Specialized Service is also true for 2026. I think that's fair to say. We have invested, as you know, in business development and AI capabilities on our Core Services, and we do expect a successive increasing momentum on our Core Services throughout the year, but we don't give different guidance.

Maybe on the impairment, Olivier?

On the impairment, so of course, we are going to continue to look at business plan for all the business. There is no, I would say, decision that has been made for 2026, as you can imagine. So we are going to look that very precisely. We will be very, very careful as we have always been in our business. But so far today, we have no specific reason to change what we have done in 2025. What we've done in 2025 was just to be on the safe side on PSG and to a lesser extent, on Health Advocate. That's it. It's not a big amount compared to the balance sheet of the group where we have EUR 4 billion of goodwill and EUR 2 billion of intangible assets. But we saw that in accordance with auditor, it was more careful to take this stance.

It was great working with you, Thomas and Olivier over the years. Wish you all the best for the future.

The next question comes from Remi Grenu from Morgan Stanley.

A few questions on my side as well. So the first one is on the organic growth guidance. Can you help us understand what you mean with a softer performance expected in Q1 based on any details on current trading discussion with clients? How should we expect that organic growth in Q1 versus the 0% to 2% for the full year?

The second one is on your cost saving plan. So EUR 70 million to EUR 90 million of restructuring costs this year. But can you help us understand the net impact if we integrate the savings that you expect to generate as soon as 2026? And overall, a discussion on the payback that you expect on the EUR 70 million to EUR 90 million you're investing in restructuring?

And the last one is probably a bit of a broader question, a lot to impact from the announcement tonight. So what do you think are the top priority for the group? Is it about first setting the right perimeter to do the divestment and potential M&A of delivering on the cost saving program, detailing the capital allocation, there's still a little bit of an uncertainty there? So just want to understand in your mind, what's the top priority in which order to understand when things are going to materialize?

So I would start and hand over to Olivier, but I would ask for forgiveness that as we speak today, Jorge is still employed by McKinsey & Company. He will start with the group on March 16. We will be then available, myself and him, to go and talk to investors, obviously. But till then, he cannot speak for the group. And so I try to cover your question.

First, guidance, yes, as we indicated in the press release, we expect based on the start of the year, we see, in particular, weakness in onshore markets. There is an increasing momentum for offshore and certain uncertainty with some clients to be below the guidance range, meaning below 0% for Q1. To be very clear, there's also the weakness with Specialized Services, but we expect for the group to be below 1% and then in continued and sustained improvement throughout the year to reach the guidance range.

Second, on cost impact, we do -- I think we also stipulate this in the press release, of course, subject to negotiation with the employee representations subject to the implementation of some of our internal initiative measures, D&I deployment, et cetera. But we expect from the EUR 100 million plus savings this year, around EUR 50 million to materialize. And Olivier can give you more details what's the net effect will be for this year also on the cash side. But we expect from the EUR 100 million plus EUR 50 million to realize this year.

And then you talked about capital allocation. I think also there, we had the discussion today in the Board. It is noted very well the request from our shareholders or for some shareholders who reached out to have an increased capital allocation by the group, and it will be considered going forward, obviously in strong collaboration with the management.

And in terms of priorities, the good thing with TP, as you know, all of the things that you mentioned at the same time. So yes, of course, there is a strong focus on the existing business. There will be a strong focus on the transformation of the group. There will be, at the same time, that's why articulate a strong focus on the portfolio review. I do believe we act from a position of strength giving the situation we are in, but there is a moment of transformation for the group that is clear, and there will be not the luxury to focus only on one thing.

Olivier?

Just to comment on the saving plan. Of course, the impact in 2026 will at best neutral. We have launched all these savings plan early this year, notably in domestic market in Europe. And we do believe that depending on what size at what speed this plan will be developing. We do believe that it will be neutral at best in 2026. And I'm sure you have noticed that we have announced flat margin in 2026 versus last -- versus 2025. That shows that we are reasonably confident that to deliver these savings. Of course, the main positive impact will be seen much more in '27 and onwards in 2026. All the job of the team today is just to make sure that we have no negative impact in 2026, which I believe we will be able to do.

Next question, please.

The next question comes from Karl Green from RBC.

I appreciate Jorge can't speak on behalf of the company or anything to do with Teleperformance, but would it be possible for him to give any kind of broad view around the market potentially just in terms of how he potentially thinks about outsourcing unfolding organic consolidation in the market? That would be the first potential question.

And then just in terms of more sort of technical questions. I think, Olivier, that you mentioned that the margin guidance does include an assumed negative impact from further U.S. dollar depreciation year-on-year. I just wondered if you could very simply just quantify roughly how many basis points of FX headwind are embedded in that flat margin guidance or stable margin guidance?

And then a final, again, margin question would be just, again, you've indicated that you would expect the specialized services margin to improve further in '26. Any kind of quantification around that would be really helpful?

Let's start with the market, Jorge.

Excellent. I'll start with the market with just an overall expert view, not at all speaking on behalf of Teleperformance, as I mentioned before, and Thomas reiterated, I will be officially with the company starting March 16. But if I look at the market and what we are seeing today in terms of trends, there's definitely a component of the rise of the hybrid workforce. And this means just having AI and humans interacting together. Sometimes AI managing end-to-end interactions and many times AI augmenting the humans to deliver a better customer experience. So I would put that on the table as one big element that we're seeing because it informs some of the other implications.

The second one that I see is, there are many companies out there right now offering their AI solutions. And some people talk about an AI bubble. Some people talk about like, hey, what is going to happen with all these companies. And I am confident that the companies that will win in that space will be the companies that have some sort of differentiation, not only from a technology perspective but also from a data perspective and the ability to integrate the solution vis-a-vis the humans.

If we play forward the movie and we believe that in the doomsday scenario, customer care will become just a bunch of models that are owned by a software company. That is highly unlikely, and we would see then many companies returning to some sort of differentiation in their customer experience strategy that involves a combination of both AI and human. And I think that, that part is something that we will need to continue tracking and seeing how it unfolds.

And then I think a little bit over your question was going is in this space, in this market, how do we see outsourcing versus moving more operations in-house? And look, right now, the market, the data that we have from external analysts is showing a slight increase in terms of outsourcing. We still believe roughly that 65%, 66% of the capacity is still in-house. So there is still ample space for growth when it comes to outsourcing. And I think that companies will be looking more and more for partners that can deliver not only on the geographic footprint but also on some of the technology solutions, the risk and compliance, the data security as they continue to do that. So that's hopefully as much as I can share right now, but a little bit on the perspective on the market.

Coming on the margin and the impact on dollar on 2026. I must confess it's a little more complex than the pure dollar because as mentioned by Thomas a minute ago, it's not only the dollar, it's dollar linked -- currently linked to the dollar, including Indian rupee, Philippine pesos and the mix of this currency versus the previous year. So what is difficult today is to predict this mix. So today, we have not a huge impact on the dollar, on the guidance on the dollar and linked -- currency linked to dollar impact in 2026 margin. There is a limited impact depending, of course, of the mix that might change. So we will update you. Probably people will be after me will update you about that because it's too early to tell.

On the margin on Specialized Service, what we can say is that I'm not waiting a big change versus 2025, except that we'll probably be better in Q1 versus last year. You remember that in Q1 last year, we have been, I must say, amazed by the impact of the reduction of the growth that we were waiting for. So now we are absolutely ready to do that. So we will be able to pass on this Q1 that was difficult last year in terms of margin. So probably a little bit better in margin in Specialized Service, everything equal, which is not going to happen, I'm sure.

But maybe as a reference, as we also indicated in our press release and the presentation, the EBITDA margin or the stable EBITDA margin guidance assumes a EUR 1.20 dollar exchange rate.

What I would say is that, of course, there are uncertainties, and you understood that. But what I would draw as a lesson from 2025 is the ability of this group across the board across a different division, across a different country to adjust quickly. Of course, it's not -- it's easier in some geographies than in others. We have been able to adjust it of course, easily in U.S., easily in India, easily in Philippines. It's more complex in domestic European market than other markets.

But what you have to keep in mind that decisions are taken quickly. They are made thoroughly quickly and implement quickly in the country, and I'm convinced that the company will continue to deliver such a reaction in case of issues or specific topic. This is something that I want to highlight because we have systems that enable us to detect quickly what's happening on the field and to react if needed, as quick as possible. Of course, there are limits to adjust, but the company is able to do so.

Maybe one last quick question in the interest of time, if there's any.

The next question comes from Nicole Manion from UBS.

I do have a few, but I'll try to be quick. The first one is just on the revenue outlook actually. So sorry to kind of go back there. But given the Visa exit should be fully annualized at least for the most part and your comments about Q1 and the growth outlook in general, the implication there is probably that the LLS situation is still deteriorating. So any kind of detail on that you can give will be great.

Secondly, just on Trust & Safety, which I think was 8% of group revenue this year. That's down from, I think, 10% in the presentation last year, which obviously is a bit of a significant drop year-over-year. I know we've all seen the headlines about some of the companies in that space maybe scaling back some of the services. But I wonder if you could maybe talk about whether it is that that's driving the step down in your numbers or whether it's AI disruption or anything else?

And then finally, just a very quick one on the onetime costs. You've indicated EUR 70 million to EUR 90 million for '26. But then you've talked about EUR 56 million of costs so far from measures that were launched starting January. Is that correct to think about that EUR 56 million as sort of relative to that EUR 70 million to EUR 90 million guide? Because obviously, that's already quite a significant chunk of that budget. So it's quite front-end weighted, if that's the case.

Okay. Let's get started. So yes, the announcement and as you see in our annual results of the EUR 56 million, all the announced social plans already today. So these are sort of earmarked in our annual results and is part of the EUR 70 million to EUR 90 million.

On second question, Trust & Safety, we do see effects, as you rightly said, for some of our clients, and it's also linked to increased automation and NI improvements in that space. So as I indicated before, there is some automation happening with this we called out a bit now what is really data services in that category. Remember, it was split between other and Trust & Safety, but it is also automation that we see in the Trust & Safety space, and that's why it's reducing. On revenue development and LLS. So we don't call out, in particular, the development on LLS or revenue. But if you look in the news and the situation in the U.S., I think you have an idea that it was not such an easy start for LLS this year.

Anything to add, Olivier?

No, no. But it's far from being a collapse. Just to be clear. Of course, what's happening on the political stuff doesn't help. On top of that, the weather did help as well, but we are not in a disaster mode far from it. I just wanted to mention it.

I see there is one last question. Maybe we squeeze that in, even though we are a little bit over the time.

From Deutsche Bank.

The next question comes from Ben Wild from Deutsche Bank.

I've got 2 questions, please. The first is on the guidance and particularly the gap between your adjusted EBIT and your FCF guide. So the guide obviously implies adjusted EBIT close to flat or modestly up before FX and your FCF guide implies free cash flow down 9% year-on-year. Can you help us understand what's going on in '26 that the results in that differential? Is there working capital reversal or any other one-off effect in '25 that reverses next year on the free cash flow?

The second question, just very, very broadly, your valuation is implying an existential trajectory for the group over the midterm. I suppose, very simply, you talked about the investment opportunities and potential divestments. But more broadly, how do you think about the relative returns of deploying capital organically in the group through OpEx and CapEx, inorganically through M&A versus returning the significant cash that you generate to your shareholders?

So I'll start with the second part and then hand over to Olivier.

No, there is nothing either in terms of working cap or CapEx or tax to be paid. I just wanted to say that we know that a significant part of our cash flow is coming from Americas. Of course, there is a lag between the EBIT and the cash items. So this is mostly the lag between the working cap that is balance sheet as of today that will be paid in 2026. So the same for the tax. But there is no specific impact we might say that we are careful as always and there is uncertainties that lead us to -- just to be on the safe side on top of that.

And the question was on...

Valuation.

So as we -- I think, at this point in time, with the new CEO coming in, I cannot say more than what we have written in the press release. The Board has acknowledged the request from shareholders also for an increased return, and we look into this. So at this point, I've asked for your understanding, I don't want to preempt any decisions being made by the new management on that front.

Olivier, if I may just quickly follow up on the FCF as a clarification point. Does the adjusted FCF include the nonrecurring restructuring costs that you've talked about in the release today or?

Yes, of course.

And then I thank you also, everybody, for your attention and your interest. I'm sure there are more questions in the weeks ahead. We're looking forward to answer them. Again, welcome to the group, Jorge. It's a pleasure to have you on board, and thank you, Olivier, for all the time, and thank you for your interest and continued support of the company. Thank you very much.

Thank you to all.

Thank you.

Welcome to TP Third Quarter 2025 Revenue Conference Call. [Operator Instructions] Now I will hand the conference over to the speakers, Thomas Mackenbrock, Deputy CEO; and Olivier Rigaudy, Deputy CEO and Group CFO. Please go ahead.

Thank you, and a warm welcome from our side. Olivier and myself are very happy to present to you our Q3 numbers. Let's have a start. I will give you a short update on the key highlights. Olivier will deep dive on the composition of our Q3 and 9-month revenue, and then we provide also some further explanation on the outlook that we're currently seeing.

Let's have a look at the key highlights. So here, 3 messages. First of all, the group is resilient, and we have demonstrated in Q3 the resilience that we have shown in the first half of the year. It is very simple. We have grown in Q3 1.5% like-for-like, which is exactly the same number that we have delivered for the first 9 months of this year, 1.5% like-for-like growth. And also in the first 6 months, if you remember, the growth was 1.5% like-for-like. So we are now at EUR 7.6 billion.

We have, and that's the second metric, also in Q3, a story of 2 tails. Our Core business, our Core BPO business is growing healthy and demonstrates strong momentum also in Q3. You see the numbers here. We are close to 4% growth in Q3, which leads to growth of more than 3% for the first 9 months. We continue to see strong momentum in EMEA, APAC and which is very positive to note in Q3 is a ramp-up and acceleration of the momentum in the Americas. You see this year in the first half year, we were less than 1% growth, and we have now demonstrated with the team, in particular with the strong performance in India and LatAm, but also an improvement in our U.S. domestic market to go back there of a growth rate of more than 2% in Q3.

We see on top of that for our core BPO services, double-digit growth in back office-related task as well as in all our AI-enabled solutions, including AI data services. So we are in this part, fully on track with the execution of our strategy.

On the other hand, Specialized Services. We talked about this in depth in our half year numbers. There, we see a different picture. As you can see, on a like-for-like basis, if we exclude the famous nonrenewal of our Visa application contract, the growth is at 2.6%. But if you adjust and if you see the number on a true like-for-like basis, we are down at minus 8.7%. And if you just look at Q3, it's down by minus 12%. We are doing a bunch of measures of addressing this. It is a highly volatile U.S. market that the Specialized Services team is facing, but they have implemented strong measures to protect profitability, and there, we are on a good path.

Third message is beyond the numbers. We are making good progress on executing our strategy, Future Forward that we presented end of June. We have set up now an institutionalized group value creation office that will focus and is focusing and supporting all our management teams across the world on 3 things: one, accelerating growth; [indiscernible]; three, driving efficiencies and to accelerate the transformation.

We are working hard on expanding our AI solutions portfolio, and I will give you later a sneak preview on the current status and more is to follow in the months ahead. And we are continuing to leveraging any in our own operation and see there are opportunities to drive process excellence as well as enhanced efficiencies.

Based on the current momentum and the latest forecast and also the volatile development in the world, we also have to adjust our financial objectives for 2025. We don't see an acceleration in Q4 that would have allowed us to reach the 2% guidance on the revenue growth side. And thereby, we're seeing now a growth for the year between 1% and 2%, a recurring EBITDA margin between 14.7% and 15% at constant exchange rates and a cash flow generation that is a bit lower at around EUR 900 million, excluding nonrecurring items.

Let's have a look at the different elements. If we start on the business side, as you can see here, the split, Specialized Services and the Core business, 15%, 85% is unchanged. What is interesting to note is the strong performance of almost 5% of our core BPO business in EMEA-APAC. This is true for a number of regions. I just highlight here a few. The U.K., our sub-Saharan Africa, APAC, our multilingual hubs, Egypt, Turkey, the Middle East really demonstrate strong performance, and we are seeing there good momentum across a number of verticals.

In the Americas, as I said, it's important to note the acceleration that we've seen in Q3. We see strong momentum in India, even moving businesses -- more businesses for our Indian operation as well as in LatAm on the domestic side as well as on the new and offshore side.

Specialized Services, as mentioned before, is impacted negatively here in the 9 months by a bit more than 8% like-for-like growth. But we see really as a resilience in LanguageLine on the profitability side, even though we see remaining challenges for the rest of the year.

To put this into perspective and particular on the BPO businesses where we spent last time a lot of discussing the different elements of growth, you can see here over the last 2 years, a continuous momentum of the business, and we see this really as a strength in our ability to adapt our business portfolio to the changing environment.

Strategy. As announced in June, our Capital Markets Day, we're in full force of implementing TP Future Forward, building a better TP for the future, enhancing our capabilities on transformation, driving efficiencies and accelerating growth for selected areas.

On the transformation side, as I said before, a few updates. We are growing double digit in these key areas: technology, consulting, back-office solution, AI data services in particular. We have launched more than 400 new AI projects in the first 9 months of this year, of which more than 150 in Q3. The whole transformation teams are busy in developing, implementing on client side, while at the same time, we are investing in our FAB solution, our foundational AI backbone solution, and we focus on driving outcomes for the clients. And that means driving industry-specific AI solutions built on our deep domain expertise and our close proximity to our clients. And on the other hand, we're building horizontal functional solution that address specific needs of our clients.

On the other side, on the flip side, if you will, this is client faces, but on the internal side, we also have launched several projects to enhance the usage of AI in TP's internal processes. As you can imagine, core processes like talent acquisition, training, coaching, workforce management, quality management are benefiting from an enhanced use of AI. And we started their company-wide program to accelerate the adoption. In parallel and closely linked to that, we do a thorough review of TP's processes and structures and see how we can, with an accelerated AI adoption, drive efficiencies for the future while at the same time, improving our operational excellence.

As promised to give you a quick sneak preview what is happening on the AI acceleration side. So this is the picture we discussed end of June, building our framework and orchestration platform for AI solution. It's really how will TP look like. And it's a combination of Agentic AI driving outcomes while combining it with the human ingenuity and the human competence that we have in the organization. And we believe we will be successful if we drive tangible outcomes for the clients. So it's not an abstract concept, but we're building and are deploying AI solution on our client operation that drive these outcomes.

For us, it's a question about being close to the client and having specific solution that works in enterprise environments. And as you can see, we have structured it in 2 dimensions: dimension one, how do we provide AI competence for functional needs; dimension 2, how do we drive point solutions with AI for industry-specific needs.

As we speak today, because TP has a rich portfolio already of AI micro services that already are built upon these FAB solutions, and we're enhancing those, have more than 400 live deployments of FAB solutions in place with our clients, and we are now honing and developing further these solutions as we move forward. And as this is really the future of TP, we will present now on a regular basis the progress along these different solution suites.

To give you maybe one example, what is live today and where we spent already a lot of efforts over the last years, it's now our new platform, FAB Assist, which combines a whole set of AI solutions TP has developed over the last 2 years. It's about augmenting human talent with AI when it comes to knowledge management, coaching, real-time feedback, quality, accent neutralization, language translation, decision support and expected action. And this is live today in practice with more than 190 clients, driving real outcomes every day for our clients.

Another sneak preview is FAB Collect. As you know, we do provide for many clients around the world collection services with human. Starting in the summer, we have developed a hybrid solution, combining human talent with Agentic AI to drive better outcomes on the collection process. I'm very proud to announce that we have already deployed now 3 client solution with FAB Collect, which is an autonomous AI voice agent for early collections, driving, of course, all the benefits also in collection services, and we have great plans to accelerate FAB Collect and the other solution as we move forward.

So this is it, the sneak preview. I now hand over to my dear friend, Olivier, who will provide some further details on the numbers. Over to you, Olivier.

Thank you, Thomas. Hello, everyone. I'm really happy to present you the figure of the Q3 and the 9 months. So let's start with this first figure. So as you can see, Q3 is a good quarter. We have announced a growth of 1.5%, which was significantly above what the market was waiting today. So we are in a growth mode, and we are going to see how it's going to be built.

Clearly, we have a positive momentum in Core Services all along the period, which is noticeable, while having a volatile business environment in the U.S., which has impacted interpretation and also recruitment process outsourcing activity. Let's go in detail to understand what happened in the 9 months.

As you can see here on the slide, you have the impact of the different, I would say, events that had an impact on the 9 months figure. Of course, the first one is the currency effect, which is significantly high and higher than that we have known over the last months. EUR 225 million. I'm sure you remember that in the first half -- in the second quarter of the year, we had already a first impact.

In Q3, we had an impact of close -- more than EUR 100 million -- EUR 104 million, which is, of course, mostly linked to dollar for EUR 50 million, but also from some other currency, including Indian rupee and to a lesser extent, Colombian pesos and Turkish lira. If you take out this EUR 225 million currency effect with a limited impact of hyperinflation, as you can see, which is negative by 0.1%, we had a like-for-like growth, which is EUR 108 million, mostly coming from Core Services. I'll come back in a minute to that. And of course, the change in scope and consolidation, which is mostly driven by ZP acquisition that was consolidated early 2025, February 1st -- 1st of February 2025 and to a lesser extent, Agents Only that came in the group in June last year -- this year.

Of course, this EUR 108 million growth is split in 2, as mentioned by Thomas. There is a significant growth coming from the Core Services, which is roughly EUR 200 million, I'll come back later on; and the negative impact from the Specialized Services, which is mostly due to the nonrenewal of the significant Visa application contract in U.S. -- in U.K., sorry. Of course, this is a global picture.

Let's go in detail on the figure. As you can see, you have here laid out on this slide, the revenue by activity for the quarter and for the 9 months. What we see, and I just wanted to highlight 2 points. The growth of the Core Services, plus 3.9% in Q3, which is clearly the most -- higher figures that we achieved over the last quarters, driven not only by the growth in Europe, which is renewed some times, 5.1%, but also with the growth that is coming back in Americas, 2.4%. Of course, the impact of the Specialized Services is minus 12.3%, which if you exclude this [indiscernible] contract, is still positive at 1.7%. You have also for the 9 months, the figures that are shown here, plus 1% for Americas and minus 8.7% for Specialized Services, but Core services is growing at 3.2%.

So if I want to summarize this quarter is what we see is a continuous growth of the Core Services, mostly driven by Europe -- continuing Europe performance, but also by the fact that Americas are back on track in terms of growth.

Now if you want to look where does come from this growth, of course, the fact that TP benefit from -- enjoy a diversified client portfolio is also really a strength and adds a value. We had a very good momentum this quarter, but since the beginning of the year in the public sector, in the media and entertainment and gaming and to a lesser extent, in FMCG and travel and hospitality. So when you look at this broad vertical approach, it helps the group to be resilient, what could happen.

And as mentioned by Thomas a minute ago, if you look where we are, care is always 54 -- as before, 55% of the business. But what we see is a growth of BPO, back-office service solution and the trust and safety and technical support solution. So we have a double-digit growth in back-office BPO solution and AI-powered solution everywhere across the region, and we have a care that is roughly at the same level that we had before.

So that are the figure for the Q3 and for the 9 months. So I hand it over to Thomas to give you an updated approach on the target for the full year.

Sure. So as discussed before, we have updated our objectives because we are a bit cautious for Q4. And the reasons are threefold. Number one, we don't expect a stronger rebound of our Specialized Services business in Q4. Yes, there's the effect of the UKVI where Q2 and Q3 have been the biggest impact, but we don't expect that the other businesses are growing more than anticipated before.

Second effect, the strong performance of our Core Services that you've seen throughout this year, we don't believe that this will overcompensate the weakness in our Specialized Services in Q4. We see in really this volatile environment, some cautiousness from our clients based on the latest business forecast. And thereby, we don't see that this can accelerate more than anticipated this weakness.

And thirdly, even from a macro environment, as you can see almost every day in the news, we had yesterday a hurricane in the Philippines that impacted our sites. We had 4 weeks ago, an earthquake in the Philippines. We had a hurricane in Jamaica that impacted our operations in Jamaica. We have the U.S. government shutdown that obviously impacts also our government services operation in the U.S., but also general consumer sentiment.

So we are a bit cautious when it comes to Q4. It's still very much sort of in line in terms of range, what we have delivered for the first 9 months of this year, but we want to give here this warning, and this impacts also our EBITDA guidance and to some extent, the net free cash flow guidance.

In context, and many of you know this, Q4 is always typically the strongest month of the year for our BPO operations, but it's also the most volatile quarter of the year that gives a little bit of uncertainty as we have for many clients, their peak period in this period. And this will ultimately land how we then deliver now November, December in the numbers. If you look at a quarterly revenue of EUR 2.5 billion, you see also the magnitude we are operating. Lastly, and just as a reminder, our financial ambitions, as we highlighted in our strategy in June are unchanged.

With this, I would open the floor for Q&A, and I'm sure you have many questions that we are happy to answer.

[Operator Instructions]

The next question comes from Remi Grenu from Morgan Stanley.

Three questions on my side, please. So the first one would be on your revised margin guidance. So can you first explain what has changed between July and now to explain that 30 bps downgrade at the midpoint? If there is any competitive or pricing pressure or if it's only in the context of what you were saying about expectations for Q4? It would be great also to have an idea of what you expect will be the FX impact. We need to add to that constant currency guidance at current level of FX, so we can kind of make the forecast on the back of that.

The second question is on the downgrade to free cash flow generation as well. Is it purely about lower organic growth and margin guidance? Or is there also some drivers within CapEx intensity or working capital that we need to have in mind and which would have changed?

And then the third one, I think you're referring to some kind of strategic review in the press release. So -- and that -- you will announce the conclusions from that in February. So just wanted to have your initial thoughts on what could be included in that plan? Is it about a review of the perimeter, cost cutting, headcount reduction? What are basically the options you're looking at?

Sure. You want to go ahead first with the cash flow and EBITDA?

Yes. Let's go with the revised guidance, which is mostly due to the fact that we are cautious, as our clients are cautious for the next 3 months to go. So there is nothing major difference from that, except this one. That's the first point. FX impact, we don't see something very different that we had in the first half, something between 20 to 30 basis points. It's difficult to tell today, but that's probably the magnitude of the topic.

As far as cash flow is concerned, I would say for a group like us, which is so big, it's challenging to forecast on a monthly basis the cash flow, specifically with our global footprint. You may remember that in the first half, we published a first half year that was down versus previous year on the back of a few one-off effects. We are catching up this effect and resuming a solid path of cash flow generation.

I do believe that since July -- I'm sure that since July, we have made significant progress in cash flow and specifically in the last 2 months, specifically on working cap and also on CapEx. So I'm quite confident. But on the back of the slight revision of our revenue and recurring EBITDA margin guidance, we want to be on the safe side, and we are, of course, cautious on that. It's around. It could be a little more, a little less, difficult to predict EUR 20 million to EUR 30 million on the cash flow of EUR 1 billion.

Remember that we have EUR 2.5 billion, I would say, account receivable on the balance sheet and missing EUR 20 million payment or EUR 10 million payment, it's difficult to predict at this level. So we are on a cautious approach, but we should be delivering the figures that Thomas mentioned a minute ago.

To answer your last question regarding transformation, we're building a future company that is the best possible partner for our clients on the transformation side, while at the same time, using the same competence and focus on driving transformation of our internal processes. As we outlined in our strategy, the second part of the coin, how do we drive transformation internally is critical. You've seen probably many news also from our industry and competitors, how they are adopting AI to drive efficiency, and we do the same process.

We look step by step, process by process, structure by structure, how we can be better set up for the future and driving this with AI adoption, but not just that, but also rethinking how we operate. We do this process now. And once we've completed it, we will present the finding at the end of February. But this is very common. I mean, if you just look in the news as some of the announcements. For all our process, our principle is to be client #1. So if we use AI to drive better outcomes, we also have to do it for ourselves.

Okay. Understood. Just one more question to follow up on that. On the capital allocation, given the share price performance -- the recent share price performance, are you thinking about or considering potentially shifting a little bit the capital allocation? And what about like increasing the level of share buyback? I think there was a change in wording in that sense with the H1 duplication. So any update on the thought process around that?

No. As we said before, this is being discussed on the TP Board level. And once the decision is being made, will be announced. To be honest, Q3 would not be the right timing, but once we have the cash flow numbers then in February.

Anyway, it's much more a 2026 decision than a 2025 one.

The next question comes from Suhasini Varanasi from Goldman Sachs.

Just one for me, please, as a follow-up to the previous one on margins. Can you please help us understand what you've assumed in the EBIT margin by division in the new outlook for the year? Now you did 12.4% margins in Core Services last year, Specialized Services was 30%. It fell a little bit in Specialized Services in the first half, but improved maybe. So how do you think about the progress in the second half of the year and for the full year? Are you expecting further deterioration in Specialized Services and maybe further modest expansion in Core Services?

Difficult to tell today. We expect, of course, a better -- a little bit better margin on the Specialized as a percentage. There is a mix effect that will have an impact because we have less business coming versus the total coming from Specialized Services. And we are going to see that in a much more precise way. But this is -- depends -- what is difficult is to predict exactly the FX impact on the Core Services where it's much more important, but probably it's where we could have an impact there. That's what I can tell you. But I won't give you precise figure at this stage. It's too early to tell.

The next question comes from Karl Green from RBC Capital Markets.

Just 2 remaining questions from me. Firstly, could you just elaborate a little bit more about the recruitment process outsourcing trends and how that's tracked through the year? And then what measures -- I mean you've indicated there's been a change of CEO for that division. What measures specifically you think could get that back to a better place in 2026? And then secondly, just in terms of the Majorel integration and the synergies flowing from that, I mean, clearly, very difficult to track in the Q3 revenue update. But just any kind of color as to what's going on there and how come that's not really sufficient to offset the mix pressures on the margin that you've outlined just before?

Start with recruiting?

Recruiting. We saw, of course, that the activity of outsourcing recruitment has been, in U.S., was under pressure. And of course, it has an impact in the second part of the year. So all the decisions that will be made has been already made to reduce the cost as much as we can to cover that. There are plenty of new initiatives that are made to offer new product. I'm thinking to a product that use AI to automate the AI process recruitment that seems to be promising. It's just starting as we speak. It's too early to give you good precise announcement and figures on it, but it seems that it's well, I would say, recognized by the market and it starts to be better. So that is the first one.

On the Majorel integration, I must say that most of the integration has been done, whether it's organic organization and merger process orientation and everything has been done, and we are -- this is behind us. There are still minor costs to be incurred in 2025 in the second part of the year, but most of it is done as we speak.

What's interesting is to note that PSG also undergoes a transformation in itself. And of course, the core recruiting services business is under pressure this year, but they have also built AI solution that address that need. Olivier mentioned it, Anna AI is exactly the solution. And when I look at the pipeline, I'm, I would say, cautiously optimistic for next year. But of course, it doesn't help with the numbers this year. So there is a headwind, unfortunately, also when it comes to the results this year on the existing business of PSG, but they're also building a new business for the future.

And when it comes to Majorel, really from an operational standpoint, the integration is over. There was some on the IT side that has happened, but it's really sort of focusing now on the future. And we will, of course, share the detailed breakdown on the EBITDA development in our full year numbers.

Okay. That's helpful. So just to clarify then, I think, Olivier, you said that the Specialized Services, that you'd expect the margins to be a little bit better year-on-year because of the cost measures you take, particularly in LLS. So just to clarify, the group margin downgrade is essentially purely a function of the mix away from Specialized Services towards Core Services?

Mostly, yes.

The next question comes from Will Kirkness from Bernstein.

Two questions, please. Firstly, I just wondered if you could just talk about Core Services and whether you're seeing any change in trends with regard to gross wins versus the outsourcing or automation headwinds? And then secondly, just on pricing, are you seeing any headwinds or difficult conversations with customers as you bring in AI either as your own solutions or in terms of processes that make things more efficient? Essentially, are they looking to take some share of AI efficiencies?

So the core trends in the Core BPO services are unchanged. As you said, automation and AI is happening, has happened in the past, and we see this every day in the business. So there's no change there. You also see a push for efficiency. And I would say it's less so on the pricing side -- of course, this also exists, but it's more on the geo-shoring side. So we see clients that look for more efficiencies, in particular, global clients, and they ask then for a different delivery location.

So we have this trend. So we see the volume stays in the business, but it moves from a, let's say, mid-priced or higher-priced location to a lower-priced location. That has, of course, a deflationary impact on our revenue, but has at least in percentage a positive impact on the margin. It depends a little bit on the location and the transition cost.

But we see this push. There's definitely a push for efficiency from -- in particularly our global international clients, and we expect to continue this in the future. That explains also to some extent, this very strong momentum that we see in India because India is a massive delivery location when it comes to efficient solution as well as back-office solution and integrated tech solution. The trends that we see and that also remains the case is really also continuous consolidation that clients want to work with reliable and strong partners. And I think there, we, as TP are very well positioned also going forward.

Did that answer your question?

Yes. There's just a question on pricing, whether customers are pushing back on sharing some of the efficiency gains around...

That's actually an interesting field. So we do have for our AI solution, a full spectrum of pricing. When it comes to AI, it's fascinating to see. I think the market, even if the -- if you look at pure AI solution player tech companies, there's a whole spectrum of gain share models, not gain share models, bundling it with existing BPO, not existing BPO. You see it's evolving. We do believe there's an opportunity to push for more gains share models, and that's why we see in the future that we invest in the business to drive outcomes or efficiencies and share the gain share.

But it's a very fluid concept. So everybody is experimenting what's the right price on token. Do you mimic also with AI solution, a human-based pricing that is linked to time and material. Do you drive it on outcomes? We see it's changing. There's also a higher willingness for clients to discuss different pricing models. We have, in fact, invested in capabilities in a new pricing team that drives exactly this new pricing model, and it will ultimately define also the BPO services industry in the next years, how this pricing will evolve. But we see it on the positive side. I think this discussion and to move away from input-based pricing to more to output price pricing is definitely an opportunity and AI helps here massively because the dynamics change.

The next question comes from Nicole Manion from UBS.

Just one left for me, please. Just on the Visa contract loss actually, it looks like that's sort of evolved in line with what you guided to in Q3. But I remember you had some impact, I think, in Q4 of last year. So just if you could remind us a little bit of the moving parts and the expectations around the annualization of that, essentially, that would be helpful.

There is still an impact for this contract of roughly EUR 20 million in Q4, negative, of course.

Other questions?

There are no further questions at this time. So I hand the conference back to Mr. Mackenbrock for any closing remarks.

It was a pleasure to be with you again. We are looking forward to present our full year results in February. As we said, we are content with the development, in particular, of our core businesses in the first 3 quarters of this year. Also in Q2, I think they did a great job of overcompensating some of the weaknesses we see in Specialized Services. We remain cautious but vigilant for Q4 and trust us that we will drive the outcomes as much as possible to present you not just the numbers for 2025 next year, but also the update on the strategic plan and the guidance for '26 end of February next year.

Thank you, ladies and gentlemen. You can now disconnect from the call.

Welcome to TP 2025 First Half Results Conference Call. [Operator Instructions]

Now I will hand the conference over to the speakers. Thomas Mackenbrock, Deputy CEO; and Olivier Rigaudy, Deputy CEO and Group CFO. Please go ahead.

Good evening, everybody. Olivier and myself are very happy to be with you this evening and share with you our H1 2025 results, and of course, answer all your questions after the short presentation. Let's deep dive into the presentation, and let's have a look at key highlights of our business in the first half of 2025. Olivier will then give a deep dive on our financials for the year. We provide the outlook for 2025, as always, and of course, answer all open questions.

How did the first 6 month of '25 look like? It was a good half year for us. Of course, there were the challenges on the FX side. But overall, we are very pleased with the development, in particular, in our core services. If you look at the numbers, we really have a story of 2 tales. On the one hand, as you can see, our core services demonstrated strong growth of almost 3% like-for-like. And in particular, in our important EMEA APAC region, we have seen revenue like-for-like growth of almost 5%.

In particular, interesting, we see an acceleration, as you remember in our Q1 presentation, we were already pleased with the development of our core businesses, and we have demonstrated further growth momentum in Q2. We're working hard on all the AI implementation, and we really see the resiliency of our core business process services. And as you can see, 3.5% growth in Q2 versus 2.3% growth in Q1.

We have ramped up new businesses around the world, in particular in EMEA and APAC, and we have improved also our client retention in our existing business with our clients. This is really a strong asset for our business in the last 6 months. On the other hand, as also indicated in our Q1 results, we faced headwinds for our specialized services. As you can see, we have grown in specialized services on a like-for-like basis. If we exclude this famous nonrenewal of a significant visa application by 3%. If you exclude that, unfortunately, the business has constructed off by minus 7% on a like-for-like basis and as we acquired ZP 4.2% as reported.

This is, in particular, driven by the environment in the U.S. that has lessened to soften volumes for a language line. And this, of course, impacted our overall numbers. Despite the headwinds in specialized services, we have grown on a like-for-like business as a group by 1.5% and which leads to a revenue of more than EUR 5.1 billion for the year.

On the EBITDA, we see also good development. If you exclude the FX effects on a constant FX basis, we are exactly at the same EBITDA margin than in the first half 2024 of 13.9%. Of course, on a reported basis, as the euro strengthened against all major currencies we see a decline by 30 basis points and Olivier will guide you through the different driver of these businesses.

We have further implemented efficiency measures throughout the group. I will give you a little bit more details for specialized services to maintain and improve our profitability in our core businesses. And given the FX effects that we've seen, we are also updating our objectives for '25 as you see at the bottom. We do now expect given the headwinds in specialized services a like-for-like revenue growth at the lower end of the guidance. The EBITDA margin remains at 15% and 15.1% objective, but at constantly fixed currencies, and we're seeing a sustainable net free cash flow before nonrecurring items of around EUR 1 billion.

Let's dig dive deeper on the different elements. And as I said, we are particularly pleased with the development of our important core services. And you see here the quarter-over-quarter development of core services with a very nice momentum over the last 3 quarters, 3.8%, 2.3% and 3.5%.

If you deep dive now and say where does this come from? We see, in particular, the strong momentum, as I said, in Europe, Middle East and Asia Pacific, where we've seen growth on a like-for-like basis of close to 6%, 5.7% in Q2. This is a super momentum driven by the strong performance in the U.K., the Middle East, APAC, Egypt, I can go on across different industries, really ramping up of new businesses having strong momentum with existing client relationships and gives a good prospect for the future. Also, our EBITDA margin for core services has improved by 10 basis points despite the FX headwinds that I described earlier.

On the other hand, unfortunately, and we mentioned this at our Capital Markets Day in Q1, we're extremely cautious about the development of our specialized services business. Unfortunately, we have seen a further acceleration shrinkage of the growth to almost minus 12% in Q2, which is, of course, impacted by the Visa management contract and the volatile business environment for operations in the U.S. If you adjust for that, we're still growing at a low single digit, and we have a set of efficiency measures in place to preserve the margins. And as you can see, we have continuously improved the margin after the dip in Q1 and are now at the adjusted level of 28% almost for the first half year.

To give you one example on this one for language line. As we anticipated, a normal growth momentum for the business at the end of 2024, we have organized and prepared the resources accordingly. Unfortunately, as we said earlier, there was a softening of volume given the special situation in the U.S. And we had to adjust correspondingly our cost base which, of course, takes some time. And you can see that after this quarter where the costs were sort of not appropriate to the demand that we've seen, we were able to adjust the cost per minute now to even a lower level compared to '24 and thereby restore the margin for language line in the second quarter. So good progress there, but remain extremely vigilant for the development, and we remain cautious on the recovery of this business in this year.

Across the board, and it's also good to see, we see that the growth in our core services is supported across different verticals, whether this is government solution, fast-moving consumer goods, media, entertainment, gaming, retail companies. So you see a bit of a shift. Last year, we saw a strong development for our BFSI and automotive clients. In the last 6 months, we see particular strong momentum across the above-mentioned verticals, which, again, is an indicator of the resiliency of the business.

Same is true for the areas where we are growing. This is unchanged. We see despite all the AI fears good momentum in our core voice and nonvoice. And we see particularly strong growth in our strategic focus areas. On the one hand, back-office BPO, growing very nicely and strongly as well as our other targeted services, which are unfortunately here subsumed into other, whether this is data services, technology, consulting analytics at a strong growth momentum.

Let's look now after we had a quick overview on the financials on the strategic highlights that the last 6 months presented. On the one hand, our strategy. We live in times of change, and we need to be an active driver of this transformation. For that purpose, we have outlined our Future Forward strategy at our Capital Markets Day in June, with our TP.ai FAB platform, orchestrating AI and human at scale. Our expansion of TP.ai data services, notably with the acquisition of Agents Only, which is a crowd-sourcing platform to have accent to this talent for data annotation and eye training work with our new AI partnerships, Ema, Parloa & Sanas, our new AI Officer that you got to know at our Capital Markets Day, Anish and the strengthening of our capabilities, whether it's IT services, F&A and B2B sales. We believe we need to be agile and focused on the strategy and to execute on the growth momentum we see in these areas.

Secondly, yes, we talked about and we saw the headwinds of specialized services, but we strengthened this important pillar of TP also in the last 6 months, on the one hand, with the successful completion of the acquisition of ZP that performs very nicely and in line with the initial plans. The integration program is well on track, and we have with Juan Carlos, a new CEO for specialized services PAUSE driving a lot of closer cooperation between specialized services and the core business.

We are also continuing to advance and we are well ahead in progress with the integration of Majorel, obviously, with ZP. We are also with the above mentioned approval of the reorganization in France. We got the approval from the French authorities, and we continue to focus on efficiency because we need, on the one hand, to drive the transformation, invest in the business, but on the other hand, ensure efficient delivery and driving efficiency operations to maintain our margin.

Lastly, as you know, as a B2B digital service company, it's about people, process, technology and domain expertise. We continue to invest in all 4 areas over the last 6 months, and we will do so in the next 6 months to strengthen the business. To give you maybe a bit of a flavor because some of you ask you to give a little bit context, I brought this time along our 3 strategic verticals that you might remember from our future forward strategy, recent wins. So in all 3 dimensions, growing the core with AI, extending our vertical place by extending the value chain for our clients, offering back-office solutions and new service lines and capturing new opportunities in AI, we saw notable wins over the last 6 months. He's just a selection where we continue to drive the transformation of TP and being the transformation partner for our clients.

Maybe to pick just a few examples what is quite exciting. If you look on grow the core with AI, we won a very large deal with a global logistic player to deploy not just one AI solution, but a set of AI solutions for scheduling, invoicing back office support where we use translation tools, interaction tools, analytic tools that drive efficiency but also improved customer satisfaction for the client. And it's really this hybrid approach that made us win that deal.

On the extended vertical plays, I'm particularly proud we recently met the clans a very large U.S. financial service provider that we are selected to support them on their risk management for the mid-market, which is not a core CX operation, but really an extend of our value chain for this long-standing client and we're very excited about the future opportunities with them.

Another one is on the health care segment in the U.S., which is an important market where we do back-office support for broker enrollments, again, an expansion of our value chain or thirdly, new opportunity AI as you might remember from our presentation with Akash Pugalia, we're very excited about our capabilities when it comes to AI data training, and we have secured 2 important wins for some or with some of our global tech clients on training and supporting foundation model and linguistic labeling work for the AI tools, which are important wins, and we will further invest in that area.

Last but not least, for a midsized tech company, we won also some IT services years. So you see the type of services TP provides the type of extension of our capability is in place and is being executed and we follow the strategic skip along these 3 dimensions.

Before we go to the financial an update as usual on our 4 capability set. People is and remains a core component of our value proposition. We continue to train them. We continue to provide a good working environment for them and we continue to expand our offering, not just for our normal employees, but also offering on-demand workforce via Agents Only. We also continue to invest in AI deployment in our core capabilities. So not just offering AI solutions for our clients, but also deploying more and more II solution in our own operations, whether this is recruitment or quality management.

We also, as a highlight, completed more than 250 AI projects in H1. As you might remember, we were at around 80% in the first quarter, and we bring them to life because we are now really in the deployment in the operational use of AI at scale on our own tools and in our open ecosystem with our partners.

One great example on that front is Anna AI. As you can see here on the chart here, have a little deep dive on. As you know, we have a specialist company, PSG focusing on recruitment services and recruitment is also a perfect example where it's the blend of human and AI driving better outcome. Anna AI is in recruiting AI-enabled recruiting assist that allows an individual human recruiter to be 4x more productive when it comes to hires. That increases the leads through better lead management and screening of CVs and applications and also reduce the wait time for applicants.

So a great example of a solution that we offer for clients through PSG, but that we also use internally to improve the efficiency of our own processes. With this quick snapshot overview, I will hand over to my dear colleague Olivier, who will guide you through the financial numbers.

Thank you, Thomas, and good evening to all. I'm going to give you a much more deep dive information about the first half. It's not a first half easy to read, but I strongly believe it's a solid operational first half and we are going to see that in detail.

If you look to the figures, there are 2 major points that I just wanted to highlight to start with. There is a solid like-for-like growth in H1 with positive moment in core service, as mentioned by Thomas a minute ago. And we -- of course, we had an impact -- FX impact in this first half and also a volatile macroeconomic environment weighing on mainly on ALS. This is a 2 major topic you have to keep in mind.

Let's move to much more in detail to have a look to the sales. So if you look at the first half, what is interesting is that we have an encouraging commercial momentum which is coming from, of course, new AI solution. But what is interesting is to see that we have, first, a big currency effect EUR 121 million, of which EUR 114 million came in Q2. So everything changed in the Q2. Of course, the euro climate versus all the currency, we'll see that in a minute, and that has an effect on our company. Of course, you remember that we bought ZP later together early this year. So you have the change in scope of consolidation of EUR 89 million. They enter scope beginning of the year. So the like-for-like growth is EUR 72 million, which is 1.5% like-for-like. This 1.4% like-for-like is EUR 72 million are split differently.

As mentioned, there was an increase in core, which is significantly higher than the EUR 72 million because we are here in a position to deliver EUR 121 million, while we were down by EUR 51 million in specialized series, mostly coming from the new contracts that has disappeared in H1.

So to make it simple, negative effect, EPA inflation impact of 0.3% in Q2, which was, I would say, coming at the very end of the quarter, in Argentina and Turkey, I can come back on it later on. It's mostly the fact that the devaluation was higher than the index of price. And of course, acceleration in core service, consolidation of the impact of TLS and an environment in LS, which is not so easy to predict.

I just wanted to give you this slide just to show that even all of the currency in which we are running and it's a part of them because we are active in more than 80 or 85 countries. All of them are depreciation versus euro. This is the first time in my life in TP that I'm seeing that everywhere, we have a negative impact from the currency, which has impact on the group.

If we look much more in detail, sorry, in the growth momentum that we were living in this quarter. As Thomas was mentioning, the core service move by to 2.3%, sorry, to 3.5% in Q2, while the specialized moved from 2.4% to minus 2.5% to minus 11.6%. But if you correct from this service contract, we are 3.8% in Q1 and 2.2% in Q2.

So there are 2 main takeaway to have here. Core service is accelerating. And this is a key question, given the fears that AI was generating and specialized service, of course, is lower than it was, but still growing. I'm going to give you much more detail by region and by quarter. You have the detail here. Of course, there is a growth which is noticeable of close to 6% in Europe in this quarter, while America is back on track, 1.1%, again in growth.

So at the end of the day, as I told you, core and EMEA are back to growth, a significant growth. and specialized service, specifically if you take out this UKVI contract that is specific, is still growing. I know it was a fear for most of you.

Let's move now to the margin. The margin has changed. As you can see, the core service margin has grown up by 10 basis points, mostly driven by the growth that we experienced in Europe and in EMEA and Asia Pacific, while, of course, the specialized services were down by 3 points, specifically linked to this TLS impact. This is -- that's where we move from 13.9% to 13.6%. And what is interesting is to have a much more precise view on this margin. What happened, in fact, in this quarter, in this first half?

If you look precisely, you have 2 major impacts, which are clear. You have the impact that Thomas explained for the Q1 in LLS, where we were facing less demand versus what was scheduled and it was correctly scheduled, but it cost us 15 basis points in Q1, knowing that starting Q2, we have been able to correct it.

TLS and ZP is a wash, if I may say, there are roughly the same amount. And you have a mix effect on the fact that the specialized service, which is a higher margin is lower than previous year. This 40 basis points have been covered by operational improvement and operational performance across the board.

So all in all, we have been able to cover this topics that were not totally scheduled, notably LLS and the mix effect by a better delivery, either in the core service and mostly in core service. But what happened is that finally, when you take in account the FX effect on the translation effect, I'm not speaking of the transaction effect, I can come back if you are interested later on the translation effect, just moving -- translating the result in euro drives a decrease of 30 basis points that hit us in this first half.

I would say after there is a few things to say, there is less synergy costs that have been incurred in the first half. It's not really a surprise because we are close at the end of the story. And we have been able to improve our operating profit versus last year. What is different is probably the second part of the P&L, where we have a decrease of the financial result, which is mainly also due to unfavorable exchange gain and loss that was coming from mainly the Egyptian pound that was very, very favorable last year, that is no more existing today. This is mostly noncash. And you have also an income tax that is a little higher than last year that will be an improvement all versus last year.

What is interesting is that if you take out this gain and loss unfavorable, you have a stable net financing cost despite the fact that we increased net basis by EUR 500 million with the bond issues that we did early January. So the average cost of the debt has been decreased by 30 basis points in the first half versus a year versus last year.

Lastly, to finish to explain also the difficulty, I would say, to understand what is economically beyond the story, which is the front-loaded outflows we had in 2025 on the cash flow. The cash flow has been hit in '25 temporarily by different topic, lower reimbursement of VAT credit that we had in 2024. So not only we get money in '24, but some money that we are supposed to get in first half will be paid in second half, notably from Greece and India, which we are speaking of a wash of EUR 30 million to EUR 20 million.

We have a full year increase in tax, which is mainly bond in H1. There will be an additional cash tax, but it will be mostly -- most of the cost has been borne in H1. You have, of course, expense related to the cost synergy plans at EUR 20 million and increased cloud subscription and rollout of AI partnership. And on top of that, in terms of CapEx, you remember that the CapEx was planned to be at 2.3%, which is already what where we are going to be, which is a EUR 40 million increase versus last year, of which EUR 30 million has been expensed in the first half.

So the net free cash flow generation is expected to be significantly improved in H2, and this is a specific situation. If we move now to the debt just to show that what happened, we have the net free cash flow we have dividend and share buyback that has been done in the first half. I'll come back in a minute to that. You have the investment, which is a better together and the investments that we did in AI, I would say, partnership, notably and some noncash issues that lead -- that pushed up to 4.4%, 82 and with supply.

The share program -- the share buyback program of EUR 100 million that has been announced in June is mostly completed as we speak. And we continue to have our BBB rating in standalone too. That's what I wanted to tell. And I'll give back the floor to Thomas to give you a final remark before the question and answer.

Thanks, Olivier, and let's have a look at the outlook. So as you can imagine, based on the performance, we updated the outlook to reflect sort of the latest development. We do not expect a full recovery of specialized services in the second half of the year, remain extremely close to the business. Watch its performance, but we don't expect that we will rebound different in the development 8 years ago. So from that perspective, we see the revenue growth at the lower end of our initial guidance.

Secondly, as Olivier explained in detail, we have these massive FX movements, in particular in Q2 of this year. And thereby, we want to specify that the EBITDA margin of 15% to 15.1% assumes constant exchange rates. And then we were not so explicit the last time, and thereby, we want to be explicit about it because that's hard to estimate how the euro dollar and all the other currency movements will be in the next years. We do expect there are some headwinds for us.

And lastly, there's no change. We have this front-loading and some of the cash outflows in this year, but we continue to see a sustainable net cash flow of around EUR 1 billion before the nonrecurring items that Olivier laid out. Also no change in our ambitions for 28, as you've seen at our Capital Markets Day, we see the traction in these new business lines and thereby, do expect an acceleration of the growth momentum. We continue to see for new, more complex services, a possibility to expand our margin and that we see after this transformation period over the next 3 years to be at 15.5% EBITDA margin and we continue to see in our resilient business model accumulated cash flow generation over the 3 years of around EUR 3 billion.

That having said, we are looking forward to your question. And want to announce one change because for many of you, you have been in contact with our Head of Investor Relations, Quy Nguyen, who will leave the company, and we are very pleased to welcome Simon Sachs as our net -- our new Head of Investor Relations Communications to the team. So in the future, Simon will be your direct contact person for all the kinds of questions you have and of course, the rest of the management team as well.

With that, back to you and to your questions.

[Operator Instructions] The next question comes from Remi Grenu from Morgan Stanley.

A few questions on my side, if I may. So the first one is on LLS. So I think you said earlier this year that what kind of growing has the diorite but was growing mid-single digit. It seems from the numbers that you're giving today that this might have deteriorated a little bit more in Q2.

So if you can make an update on that current trading in LLS and how you're thinking about the second half of the year. You're not guiding for recovery, but equally, could there be a deterioration there. So that would be the first topic.

The second 1 one on the acceleration in core services in Europe. So can you maybe provide a bit more details there on what has driven that? And the level -- and especially the level of margin that the additional volumes are coming at especially interested in the activity coming in back office and AI data services that you're referring to?

And the third one would be on the guidance, the margin guidance at current FX -- constant FX, sorry, 15% to 15.1%. If we assume that the ForEx remains stable compared to where it is today. What would that translate into in terms of reported margin, just to understand the gap between that and what we should expect for the second half?

So let me answer the first 2, and Olivier can answer the FX effect on the profitability. On the business development in EMEA and APAC. As you remember, we have invested last year in our business development team. And we have been very strong in our ability to win new contracts, ramp new contracts and grow with existing clients in EMEA and APAC, notably in the fast-moving consumer goods, government sector, some e-commerce clients that we ramped. It's a mix of, I would say, our operation to delivery like in Egypt or Turkey, strong local business wins in the Middle East. APAC has been further good market in terms of revenue contribution in the first half of this year and also the U.K. with their delivery locations in South Africa.

So the team has really done their a fantastic job across the board of winning, growing new contracts and growing with existing clients in these 3 sectors. And we have seen also in EMEA, APAC, the benefits of the synergies. If you remember the chart, let me maybe go into this, on the detailed numbers here, you can see really in the EMEA APAC, despite some of the FX headwinds a growth of our EBITDA margin from 9.3% to 9.7%. Yes, there are some challenges in some domestic European markets, as we talked about. But this region, and it's a significant part of our business, is resilient, growing strong momentum, Africa and the other regions I mentioned for our operations.

So there, it's really it's quite exciting to see us that some of the challenges we save on the specialized services are compensated by our core business and EMEA APAC there has a strong component. Americas is a bit mix pictures. As you know, in our numbers here, we have 3 big regions part of it. North America, Latin America, Philippines and India.

India continues to be a growth driver for the group. So we're very pleased where we're for the development. North America remains challenging on certain fronts. In Latin America, we see an interesting picture. It was a little bit weaker development last year, but we see a strong development of our domestic businesses. Notably Mexico, but also Brazil, for instance, there we have the FX issue unfortunately. But it's really -- you can see it. We win new clients, we extend the services line is really encouraging to see. So if you peel the onion adjust for the FX effect. There are some challenges there, in particular, in the Americas. India is growing, LATAM is growing domestically as well as in our new offshore operations. There, it's quite encouraging to see.

Unfortunately, we still have the happens in North America. Then there was the question on specialized services. So we don't expect a recovery. Language line in particular, we don't break out the growth rates, but it's growing it's still growing, but low, low single digit, so to your point. And as you can see, obviously, for H1, we see a reduction of our margin, unfortunately, by 30 basis points. So the capability, how this year will end up, if we are very pleased or pleased or just modest will depend on our ability to continue the momentum in core services, that's really strong and positive to see. We see the pickup in back office work. And how much headwinds we have for specialized services in the second half of the year.

Thank you. Before I move to your question, just to be clear that the margin in LLS in Q2 is higher than last year in terms of percentage. So we have been able to monitor that very closely. Of course, volume is difficult to predict. As far as FX is concerned, this is as you understood, is a complex year. There are 3 topics that I just wanted to mention. First one, hyperinflation.

Frankly, I have no idea where it's going to come. My guess is that probably will revert in the second part of the year as often it happen. When the price will, I would say, adjust with the devaluation or in the other way. That is probably what it could happen, but I cannot guarantee it.

On the transaction issue, the group has covered mostly all of them. So there is no marginal impact of the transaction because we are covering roughly EUR 3 billion a year. And this is -- this was down sufficiently early in the year to cover it. On the translation issue, which is not covered, the impact was 30 basis points as we speak.

Today, we believe based on what I know, what I feel because there is no certainty, we should be roughly in this tranche. There is no guarantee. There is no guarantee it could be better, it could be worse. But it depends a lot of the euro versus all the currency. Most of the time, a lot of the currencies I mentioned it earlier on, are somewhere linked to dollar. And when it translates to euro, it has an impact significantly for us.

The next question comes from Karl Green from RBC Capital Markets.

Yes. Just a couple of questions on specialized services and then a separate question, if I can. Just in terms of that development between Q1 and Q2, excluding the TLS contract loss. So excluding that loss, like-for-like growth was 3.9% in Q1 to 2.2% in Q2. Given the very big swing in the overall like-for-like growth.

Can you be a bit more specific about the magnitude of the contract loss in euro millions in the second quarter? And how should we be thinking about that in the second half that, I think, seems to be the big gap versus consensus expectations going into this print.

The second question around this general specialized services topic is given what's happening, are you more inclined to think more deeply about potential divestments? And then the final question, completely unrelated. At the full year stage, you did move towards giving the market an adjusted EPS metric. I just wondered if you could explain the logic for not providing that on this occasion, given that there are lots of moving parts below the EBITDA line in terms of tax which you talked about FX gains, et cetera? What's the logic there?

Olivier, you want to start?

Yes. Q1, Q2, specialized. So the big impact, as you understood, was TLS and the total impact in Q3 will be again something around EUR 50 million in terms of sales. It's a big impact. The bigger impact are in Q2 and Q3. It will start to reduce in Q3. Divestment, I'm not sure we are looking to that as we speak. We are always looking to possibility of moving any solution for the shareholder. That's it. But today, we are not -- we are much more focused on the delivery than on potential move.

As far as EPS metric adjusted EPS metric are not done is mostly because it will be done at the year-end. It's not done at the first half. We should have done it, we can make it, but this is not -- this was not something done on purpose specifically.

But maybe to add the point, I do believe, yes, we do see some weakness in our specialized services right now. But these are highly valuable business. If you look at reference cases, whether it's in the visa services segment, whether it's in the interpreting segment, in our specialized segment for recruiting solutions that are now AI enabled or health advocate that we presented at the Capital Markets Day, I strongly believe these are truly valuable company assets that -- yes, there are some headwinds this year, which we manage very closely. But these are valuable assets that if you look at relevant peers have sort of -- would maintain a higher valuation.

You had another question or that's it?

The next question comes from Carl Raynsford from Berenberg.

Could see some solid momentum in the core services. So my question is actually focused elsewhere, I'm afraid, and I've got the 3. But the first, just another follow-up on LLS. I'd have thought the LLS trajectory should be sequentially improving even from Q1 to Q2. So how can we gain comfort internationalized services returning to that high single-digit or double-digit growth rate over time? You explain these are valuable assets.

But to me, given your commentary, it does sound like there could be a structural growth issue there with LLS. And you've got more real visibility as to growth actually recovering. So it'd be good to make your thoughts there. And the second, you note front-loaded outflows regarding free cash flow. But could you please quantify how free cash flow in the second half should be so much stronger to get anywhere close to that $1 billion level? Now when I look at the slides on -- you'll note, sorry, on Slide 27, it's not so clear to me. I also note you've removed the guidance for deleveraging this year, which was in place of Q1. So I presume now that you're not even sure about that.

So doesn't exactly give me a ton of confidence you're able to generate that EUR 1 billion of free cash flow. So if you could quantify that, please.

And lastly because you again quantify the level of OpEx investment in the first half, please? So of that 40 bps operational improvement gain you stated, is there extra cost investment that offsets the gain perhaps up in there, i.e., is the underlying improvement higher than 40 bps, and you've invested in workforce, let's say, would sort of offset that slightly. That be good to know.

On LLS, let me start with this and hand over to Olivier. I think you have to distinguish between structural changes in the business and temporary changes in the business. If you are -- in the U.S. right now, you see that the environment, in particular for people who are seeking the support of a language line, which are not non-English speakers are today very difficult. And this is part of the slow growth demand that we see in the business that there is less demand from these more non-English-speaking groups.

We don't see a structural shift in language line that on the interpreting solution side, the opposite is to, I think it's a very resilient business that despite the situation in the U.S. is still able to grow. So that's why if this would normalize, I don't know when this would be. We do expect this pent-up demand in this normalized demand to be significantly higher as it is today.

And it's a non-normal situation that we don't expect to be sustained for the next years. We do -- as we want to be cautious, we don't expect a recovery in the next 2 quarters. But we don't expect that this would this structurally shift over the next years.

On the free cash flow, just to be clear, the EUR 1 billion -- around EUR 1 billion is excluding nonrecurring items. You remember that there are EUR 20 million has been spent in the first half and probably there will be another EUR 50 million that will be spent in the second part of the year, linked to the French plants that has been mentioned. But why we are, I would say, so confident.

We are, I would say, between brackets, so confident. If you look what happened last year. Last year, we gained EUR 600 million of cash flow in the second part of the year. Here, you have rough between EUR 110 million and EUR 140 million that are, I would say, moving from H2 to H1 versus last year. When you put that together, I'm not saying it's granted and everything is okay.

Of course, you have this nonrecurring item that will be impacted cash flows for this year. They might probably have some FX issues is the dollar continued to slip. But we are -- what we know is that there is as always, in this company, a lot of business happening in the second part of the year, a lot of revenue, a lot of cash, a lot of margin that is happening in the second part of the year.

So it's not unprobable that we will achieve it. Of course, there are uncertainties, but we are on it, and we will be reasonably, I would say, we are reasonably, I would say, confident that we will achieve this figure.

Again, I just mentioned, there is a cost of this synergy plan that will be deducted from this around EUR 1 billion figure.

Third question.

Third question, I'm sorry. I'm not sure to have understood it, meaning that weather this 40 basis points will continue in H2. That was your question? Am I correct?

It was more just you've been pointing to OpEx investments into your people, i.e., tech -- AI-savvy people...

We had that last year, and we have a remaining impact in the first half, which is not dramatic, but existing. So we should have less issues than we had last year, notably in the second part of the year, yes.

Okay. Just a quick follow-up on the second one on sort of free cash flow, sorry, just with the fact you've removed deleveraging from your guidance. Should I read into that in any way? Do you expect leverage to go up?

No, no. I'm not expecting labor to go up. What I didn't have at that time, it's EUR 100 million share buyback that was planned. So there will be a lower deleveraging to where -- so of course, we are going to deliver to what level exactly, I don't know precisely, but I have to take out to add EUR 100 million debt following the share buyback that we did in July.

And if you have read our press release, and we did actually have our Board today, we have a very active discussion what is the best capital location for the company in future value creation. We did announce the share buyback that has now completed the EUR 100 million. And so this needs to be taken into consideration as this was an ask that is being now deliberated on the Board.

But again, again, in case you will be afraid by debt. I can tell you that the group is totally in control of this debt. Have access to liquidity, whether it's short or medium or long term, and we have no issue of financing, clearly.

[Operator Instructions]

The next question comes from Nicole Manion from UBS.

Sorry to labor the point a little bit around LLS because we can go back there. Just on that weakness that you've seen, first of all, is that something that was deteriorating through the quarter sort of month-on-month. And then maybe in addition, how much of that is being driven by I guess, you call it the political backdrop in the U.S. and demand and willingness to engage in these services and how much is related to other issues, including kind of potential AI overhangs. I know you mentioned that the high-value interpretation services are still growing, I think, low single digit, which obviously implies that everything else is seeing quite a bit worse. That was the first question.

Nicole, your question was on the cash flow or on the development on the revenue side?

On the revenue side, from specialized from what I understood now, am I correct?

I think really when you talk to the team, there's really...

Yes. Yes.

A very special situation right now in the U.S., I cannot say differently, and that has obviously an impact It's still remarkable to seeing that they got their costs under control. There are hundreds of individual measures in place to ensure the efficiency of the business and to maintain the growth momentum that the LanguageLine has today, I don't see significant AI impacts in the business today.

It is not -- that's why I say, it's not a structural driver of the business. It's a demand driver on the business, which puts the entire interpreting solution for this special demands in a particular case. So that's why I'm not -- that's why I don't see that in the numbers. Yes, you're right, there are some headwinds also on the other specialized services business. But the main impact is for specialized services, really the loss of the TLS contract, and the headwind versus the normal demand on language line. And as we said earlier, the team has prepared for normal double-digit growth. That's why they put the resources in place, and we had really a lot of measures to do to adjust for the resources accordingly to get the costs under control.

But if I may add, Nicole, you have to understand that a lot of people who were using our service are afraid to use them, not because they don't need it because they are afraid to be, I would say, kept by ICE and sent back to Latin America or to Mexico, even if there are, I would say, regular U.S. citizen. The fear is clearly one of the major driver of this evolution, in U.S. and U.S., of course.

Yes. That was my question basically whether...

To what extent -- to what extent...

Sort of demand weakness.

To what extent, it's difficult to figure it, but what we get as a message is people. And we were looking to figure again with Thomas recently. You see the 20th of January drop, which starting the 21st of January.

Yes, I said it really, we spend a lot of management time and attention on the development. We see daily reports on the movements on the volume we really have all our eyes on the business. The team of LanguageLine does a fantastic job. I cannot say differently. They really -- it's a super company. They deploy AI in their internal offering, how they develop it. They have a super close and great sales organization. It's an absolute superstar in the segment, but they face these unusual headwinds this year. And that's -- we have to weather through.

I was hoping we were away expecting a recovery in the second half. Which is not happening. As you can see, if you watch the news in the U.S. every day. But we see the business is intact, healthy and ready to grow again when the situation has been normalized.

And this volume are particularly sensible to what's happening, notably what happened in Los Angeles 1.5 months ago or a story like that. So we see a direct link between these 2 phenomena.

But that's why I really have full trust and the team. Simon now with JC, they do a fantastic job. And it's a healthy business that has to weather this period through. We support them in any means. We mentor them in any means, but this has, unfortunately, an impact on our business, unfortunately.

On sales, the margin has never been so good in LLS in percentage.

Other questions?

If I could just ask a quick Yes, if I could just take one second question. Just moving to TLS. Sorry, just a basic question in some ways. But could you just clarify a little bit about when the impacts started to come through here and maybe the annual sort of value of the contract you lost because you did call out an initial impact on growth from this contract loss or nonrenewal rather in Q4 of last year. And now it seems to be quite choppy through the quarters of this year. Maybe that's some seasonality, and I'm not sure. But could you clarify why that is because it's quite confusing, I think, from the outside that there was an impact in Q4 and then, obviously, an impact in Q1 and then that's actually become significantly more pronounced.

And yes, just trying to understand how we should think about that from here and on?

It's a complex situation. When you stop such contracts, you remember that TLS was supporting delivery of Visa for many, many, many countries for the United Kingdom. So there are some parts of the business that have been close or stop earlier in Q4 that some of them have continued in Q1 because the government asked us to continue for until end mid-March for some place. So this is exactly what happened. So a significant part has started to be reduced in Q4. There are still business, I would say, on a comparable basis in Q1 but significantly less and totally disappear in Q2.

Other questions?

The next question comes from Ben Wild from Deutsche Bank.

Just a few questions for me. Firstly, away from LLS and specialized services in the core services business, there's quite a big divergence in growth between the Americas and the EMEA and APAC divisions. Can you just maybe explain the divergence that's happening there because the segments are obviously very geographically diversified in terms of the geographies that they're an serving. So any color on what's happening geographically or whether it's a function of client mix is resulting in such different growth trajectory from the different subcomponents of core services.

And then a second question on Specialized Services, please. And a question -- another question on disclosure. Thomas, you mentioned that some of the businesses within specialized services have very high valuation benchmarks in the private and public markets. Are you considering giving more public disclosure on the constituent parts of the Specialized Services businesses to help us understand the moving parts within the businesses and maybe to help us value them independently?

Yes, we have -- there are discussions, but nothing has been decided and disclosed, but I think if you look at the past as some of the assets within specialized service has been acquired, you can drive your conclusions of the businesses, but these are really -- and we gave some examples at our Capital Markets Day. Specialized niche services business that target a specific client demand that have higher margins typically. And I do believe echo a higher valuation that we are currently seeing.

So that was just as a note on the business. On your different momentum question, I think it's reflected that the Americas, what I said earlier, a relevant portion of that segment is North America, notably Canada and the U.S., and we've seen a trend for many, many years, but also in this year of moving businesses onshore to offshore location. In the English-speaking world, of course, this is much easier, whether this is India, Philippines, South Africa. And this affects, obviously, the growth rate. That's why when you dissect our Americas segment in India, Philippines, Latin America, North America, you see a very different growth momentum where India is growing high single digit. And the other segments, in particular, North America, sees a decline in the top line, which is true not just fast, but for the entire industry. And so there's the relative weight.

Maybe just as a quick follow-up.

Sure.

In terms of the English-speaking world and the offshoring dynamic, is there any differing pace of AI disruption happening within the English-speaking part of the business versus non-English speaking. And can you -- is that something you're seeing or Is it purely an offshoring effect?

No. The beauty of AI, and that's why I say, it's, of course, multilingual. So that's why the language barrier that exists with AI. So we have really positive discussions with many clients across the world and whether you are a French client, German client, Spanish client, U.S. client, every Agentic AI voice spots, speaks every language. So there's no language barrier as such. And also the willingness to try things -- we have many European companies, many Asian companies for a very active discussion with us.

We're winning deals based on Agentic AI solutions by combining a hybrid offering with AI and human-led BPO together across the board. In fact, we actually had today a very recent discussion with our Head of Business Development and Sales in Europe of all the deals that you've seen. So there is no difference in the dynamic for our U.S. clients versus our European clients.

And it's good. Really, we have to embrace the and I think it allows us an opportunity to transform this business if we do this very proactively.

Maybe last question?

Last question. I don't know if there's something left in the queue.

There are no further questions at this time. So I hand the conference back to Mr. Mackenbrock for any closing comments.

We stick to this chart, maybe we'll look at it. We really are pleased with the development in our core business. The teams are a fantastic job growing like-for-like. As you can see here, 3.5% in Q2 is encouraging momentum and we're really looking forward to continue this momentum in Q3 and Q4. We are closely monitoring the headwinds in specialized services. Also there, the teams really have our full trust and doing a fantastic job. But these are headwinds that needs to be managed, that we see temporary, but we cannot give you a date for this year if this will end.

And we have, of course, on the group level, the FX headwinds that Olivier explained earlier that we continue to monitor closely. But we look with optimism and encouragement to the next 5 months and 1 day to then present to you in the meantime, our Q3 numbers and then our full year numbers. Thank you for your attention.

Thank you to all, and have a great summer.

Have a great summer. Take care. Bye-bye.

Welcome, everybody, to TP Capital Markets Day. I'm Mike Lytle, the CEO for TP in the U.S. I'm still going to call us Teleperformance a few times on accidents. So just I'm self-declaring I'm going to make that mistake.

We have a full agenda today. And so just a couple of housekeeping rules. If you'd like to ask a question during the session, there are microphones on each of the tables. Outside, you may have seen there's a few stations set up to demonstrate the technology that you'll hear about today and a real live demonstration. So you can see how the technology actually works with our clients and end customers. The TP team will also be outside. So ask us some questions, talk about the things that are on your mind, and we'll be happy to engage.

I'd love to introduce Daniel and Thomas, who will walk through the strategy overall for TP to kick off our day.

Daniel? No, video first.

[Presentation]

So good morning, everybody. Thank you for being here, and thank you for the people. I think there are 170 people who are following us remotely. So thank you for everybody. We are going to try to be efficient and clear.

Today -- first, my name is Daniel Julien, for the ones who never heard about me. I started in this company a few years ago, and I'm still the CEO of the company by chance. Today, my partner in crime and my partner in management, Thomas Mackenbrock; and myself, we are going to introduce the whole session telling you where we are, where we want to go and how we want to go there.

There is something very important because you are going to see the day is going to be very focused on the AI integration. But there is something very important for you to keep in mind, which is that the motto, the main line of Teleperformance is, AI augmenting our human experts. And why? Why? Just because all human decisions are made from 2 factors: emotions and reason, emotional and rational.

All the practice, decades of practice, all the academic work show that in a human decision, typically, at least 70% of the decisions, has an emotional component, and the rest is supported by the rational. And so our role, of course, is to have human experts who are empathetic to the problem that meet the customer, but we want to empower them with all force of the AI. And this is what we are going to try to show today.

So where are we today? And then Thomas is going to tell you where we go tomorrow. Okay. Our job is very simple. We are a business service company, and we are here to create competitive advantage for our clients. And how? Through digital business services, meaning taking full advantage of the digital.

You know we are a global leader in the customer experience management market. We are the #1. We serve 170 countries. Our revenues are over EUR 10 billion. That's the scale. We have scale and we know what we are talking about.

Second, we are trusted by our clients. In the Core Services, we have more than 1,500 clients with an incredible stickiness, an average of 13 years, meaning that mainly among our largest clients, our clients for 20 years, 25 years, I mean, 1 generation, very interesting, specifically when those companies can be among the top companies of Silicon Valley, for example. And 2/3 of the top 100 companies, of the top 100 brands, are clients of Teleperformance.

So we have scale, we have a strong relationship with our clients. Yes, we are a people service company. And being a people service company, we want our people to work in the best possible condition, intellectual condition and physical condition. That's why 97% of our people who work in subsidiaries of Teleperformance that are Great Place to Work certified.

Also, more than 70% of our promotion are internal promotion, meaning that we absolutely want our partners, our colleagues to blossom to their full potential. And the number of people who started at -- as customer expert at TP and who now are CEOs of a company or head of a large vertical is something pretty amazing.

And finally, we are a public company, and we deliver value to the shareholders. And by the way, we return quite a lot to the shareholders. In '24, our EBITDA margin was 15%. We increased the dividend per share between 2022 and 2024 by 75%. And much more importantly, today, is the fact that we created, in 2024, in the environment that you know, EUR 1 billion free cash flow. And we expect to continue.

But before to pass to Thomas, I would like to show you that. You know Everest Group. Everest Group is the business analyst that is probably the most renowned business analyst company. And this is where they placed us in the customer experience management. You'll see for the Americas and you'll see for EMEA. So this is external data.

I'm not going to mention the industry awards, but I can tell you that I was very happy 2 days ago to discover that we got a Golden Award in India for integrating AI solution in our services.

I can tell you that the company that I created 47 years ago has changed multiple times. We are like this animal who leaves the skin, changes skin. We are continuously morphing. And in fact, the pre-2000 have been the years of the geographic conquest. And this allowed us to start the geo-strategy, the offshoring process. The 2000 to 2010 was the fact to put together fragmented channels, in one single seamless experience, the omnichannel. 2010 to 2020, we started to integrate much more the efficiency of the digital with robotic process automation, with the optical character recognition, and that was a big transformation.

And then after the 2000 and now with the emergence of the transformers, of the large language model, of this fantastic possibility that is offered today, we are going to the next step, which is the cyborg customer expert, made of the human empathy and all the intelligence that we can get from the digital and the Gen AI. And of course, we are going to speak about agentic today.

So having said that, I think I'm finished, and I need Thomas to come, if he accepts to stop to speak with his friend.

Thanks. So good morning, everybody. It's a pleasure to be with you all today. And we spend the next 2 hours to talk about the future of TP. We will talk about why we are excited about the future of TP, what our plan is to deliver this exciting future, and what ingredients and sources and capabilities we already have in place today to do so. There will be plenty of information, but I would like to leave with you with these 4 messages, and we will dive through all of these messages now in the next following presentation to explain what we already do today, what are real use cases we see in our businesses, and how do we want to act and expand upon them going forward.

First is our strategy. We have a clear strategy in place to expand our business and to accelerate our growth momentum going forward. As you know, the guidance for this year was 2% to 4%, and we see clear pathways to accelerate this momentum going forward. Secondly, we're very proud to announce the platform TP.ai FAB. FAB, as you might think, stands for -- not for fabulous, it also stands for fabulous, but it stands for Foundational AI Backbone, and essentially is the platform that is built upon the hundreds of AI projects we have done in the last 2 years, our deep know-how working with our clients across multiple verticals, our own proprietary solution as well as the partner solution we want to bring in our ecosystem. And we will explain to you how we combine this agentic AI layer with human orchestration, because we believe this addresses the real pain points in the use cases that our clients ask us.

Third key message, and you will see many colleagues today on stage, is that we believe we have the right team in place. They're hungry to win, they're hungry to transform the company. When I travel around the world, I'm always amazed by the quality of people TP has in every geography, is really an all-star team. And you will see some new faces today and some old faces who will explain the journey we are having.

And lastly, the financial numbers. TP, as you all know, is an asset-light business model that is highly cash generative, and we expect this to continue this in the future. We estimate around EUR 3 billion net free cash flow in the following years. This includes our internal AI efforts and a significant portion of that will be returned to the shareholders going forward. How are we going to do that, we will guide you through.

So let's double-click on these points. First, our strategy. So I always say our business is 90% about execution, but you need to know what you want to do and what you don't want to do, and you need to know where you can focus on because you have to have a right to win in the different areas. And it's very easy. First, we see plenty of opportunities in transforming and growing our core operation. Yes, we need to change what we do, but we can build on the capabilities of people, process excellence, technology and expertise in our core business, and thereby broaden our existing clients extend our share of wallet with them and thereby drive growth. You will see multiple examples how this is happening already today and how do we want to expand on this in the future.

Secondly, and we talked about this in the last investor calls, I'm very excited about our capability to expand our value chain. What does that mean? So historically, we have been very strong on front-office and mid-office services, but we see today in our business, as you know in our numbers for '24, we see double-digit growth in our back-office related task. And you will hear from the colleagues Mamta, Miranda and Himadri how we expand our value chain from the front-office first to the back-office in our core business services as well as for our verticalized solution. And expanding display, having a more vertical view gives us more growth momentum going forward.

And thirdly, AI value chain. So with the capabilities that we have today, we see us well positioned working with our existing clients to drive new opportunities in AI data services as well as technology and consulting. We launched, as you might have seen in February, TP.ai, and we're very happy to have Anish being with us today is our new Chief AI Officer. And here, we will demonstrate how we see already today in our business the capability to capture new businesses, in particular in data services as well as in tech-as-a-service.

All of these elements combined, it's not a surprise because we operate in a large market, yes, it's changing, but it's a huge market, allow us to increase our TAM and really focus and execute on those areas we see the growth momentum. And the ability to act quickly with our capabilities that we have in place will determine success.

Second key message was TP.ai FAB. So you can say every BPO announced a new platform. What is different? What's unique? Our approach is not -- we are not a software company. We are a B2B services company, and we're listening to the needs and problems, if you will, or opportunities that our clients present us. And when we listen to them, one key topic we hear often and often again is how do you orchestrate? How do you intelligently, seamlessly, safely combine the human experts with the AI because very rarely, every tool and every news you see, every demo you see on the media is a plug and play, in particular, in enterprise environments.

So how do you build intelligently that orchestration? And I really do believe, TP, with its global footprint, its global scale, in people process and technology, is uniquely positioned to drive that orchestration. And we will show you how this is already happening today, and how do we expand this in the future.

Second key topic and message we received from our clients is safety. In order to win and to operate working for the largest companies in the world, you need to have a secure, safe, open cloud infrastructure. We have, if you have in the breakout time, our CIO, Dev Mudaliar, he will explain to you how we are building this infrastructure. We invested 3-digit millions amounts already today in our safety, and we will continue to invest in that area. And this is an integral part of our TP.ai FAB.

And thirdly, specificity. What does that mean? If you read about and if you hear about AI, it's very, very rarely that a general AI solution is able to drive real outcomes for clients. You need to pick on the value chain, on the process that you drive for a client specific use case, and you need to apply for that use case to make it practical. It's not that you use a general tool that will solve all the answers. It's about vertical-specific blueprints that drive outcomes for specific needs, whether this is claims management, subrogation, you will see some examples later.

And all of these 3 elements, we put together. Many things are already in place today. This is now the orchestration platform that we show, but we will continue to invest in it and build blueprint layer for blueprint layer in the coming months and years. All of this is built essentially on TP's expertise. As you think about it, today, TP manages already in an existing business billions of interactions every year. We have 40 years' expertise with processes. We have thousands and thousands of experts on the tech side, but also on the industry side. And this all feeds our TP.ai FAB where we will go further to the market. We have, in our next session, a detailed look to guide you through how this is live today, where we further invest. But this will be an integral part of our strategy in the coming years.

Third message, do we have the right team? So when I talk to many investors, do they say, "Yes, you're so successful over the last 40 years. Do you have the right -- not just the right resources and capabilities and clients, but the right team onboard to do that?" And this is really, for those of you who have visited a TP site who has interacted with the management, it's really part of the DNA. Its operational excellence to drive it. And that will be a key component going forward.

And as Daniel said, we are very fortunate to have great talent in-house, but also bring on new talents. And in today's presentation, we have 4 new faces that you might not have seen before. In the next session, we have Anish Mukker. Anish has had a very successful period at Genpact and Amazon. Then joined us in 2022 as the CEO of India. And India is today our largest market and really on the forefront on large-scale AI transformation. You just heard the latest award they got. And Anish agreed to take on this new role as our Chief AI Officer to really bring this life for the group. So very happy to have him.

The other person that is new, you might have seen in the past, and who's presenting together with Anish, is Akash Pugalia. Akash is a former Accenture, then Cognizant, then he joined us and built the trust and safety business. As you know, trust and safety was a new business for us, which is now over EUR 1 billion in revenues. Then he left, unfortunately, I don't know why, but we brought him back at the end of last year, and actually to drive TP.ai data services. If you hear all about the company, scale, et cetera, this is a huge opportunity for us. It's a market that's growing close to 30% a year. And Akash and the team is laser sharp focused on driving this, and we are very happy also to announce an acquisition, a technology platform, in that area today.

Third new face, you will have later, is where is see, JC. JC is also a BPO veteran for over 20 years in the industry. He looks older than he is. He worked at IBM, was then the CEO of our Latin American business. As you know, it's really a very successful business across Latin America. And he took on the role as CEO of Specialized Services, to really not just support the growth, but even to ensure a tighter integration with our existing business process services business.

And then the fourth new face you will see today, unfortunately, just by video, is Sherri Turpin. Sherri is the CEO of ZP, based out of Austin, who came through the acquisition we announced last November and closed in February, and very successfully led the business, it's a super exciting business, over the last years, and we are very happy to have her onboard.

So these people are not, of course, the whole team, but they are representative of the strength and management team that we have. We will continue to invest in the people because having the best people on the board will be critical to drive the execution. As I said, our business is not about finding something new, but executing on the strategy as well. And I do believe all of them are very hungry to win in the coming years.

And last message, as I said, how do we drive value? We are just the agents of our principals, which is you, is we want to achieve higher top line growth, we want to increase our margins, and we want to deliver cash flow to you. All of that is part of our strategy. We're seeing essentially a doubling of our growth rate post transformation and expansion of our EBIT margin as well as a continuous delivery of a significant cash flow in the years that will be used to invest in the business to further deleverage the company as well as return it to our shareholders.

That's in a nutshell. You will see all of this in detail with use cases to bring it to life. But before we do that, I just want to tell you 3 stories, because our reason of existence are essentially our clients, how we have this not as a future strategy, but as a reality today already on the ground.

The first example how do we grow the core is a global electronics company. It's been a long-standing clients. We have also Danny, where is he in the room, is working quite closely with them. And it's a great client example where we are able to drive by implementing every month, I would say, every quarter, new technology to the client, the customer satisfaction up. Actually, the cost per interaction down with the client, so the actual cost per interaction down, but we increased our revenue with the client by offering new services, new solution to the client and broadening our scope and share of wallet. It's a global business, and we continue to invest with them. You see here our real-time translation software, our customer Copilots. It's a super case where we basically orchestrate our existing human delivery with AI and thereby expanded our revenues.

Second example is a major U.S. bank. As I said, I talked about that we have really grown last year significantly in our back-office services, and this is a great example. TP is a reference for the BFSI industry. We have more than 65,000 colleagues around the world supporting banks, insurance companies. We have more than 350 clients in that industry. So really, really deep know-how. We have Mamta later, who worked in that space, presenting in more detail what we do with them. And we're really proud to work with the largest banks in the world.

And often, we start with the front-office service, a customer service, and this was also the case. But over the years, we expanded line of business -- of line of business, now managing the entire end-to-end customer life cycle, entering into fraud prevention. So you can see how we really, if you look at on the client discussion when we do our client reviews, over the years, added line of business of line of business. So we have now 12 business lines, quadruped the revenue. But really being the best strategic partner for these clients, delivering operational excellence, delivering innovation is the path for us to grow further. And this is an element where we have, as I said, grown double digit last year in back-office services. We continue to invest in that area, and we are very, very proud of the work that we have done for this client and many other clients in that space.

And last example is -- and I know that many of you also asked for this, is data services. So why do we feel we have a right to win in that space? And what do we want further? Obviously, any AI is only as good as the data that it uses. And you see almost every week in the press news about hallucination, wrong data. And this is an opportunity where AI does not replace even the wrong answer, where it creates new job opportunities because you need human-reinforced learning to make AI better, to check accuracy, safety.

So it's a whole new industry that is just about to start, really just about to start. If you look at the reports, it's considered to be single-digit billion, but growing rapidly. And if you see a future where AI will be ubiquitous, having humans checking the data, the results and creating this reinforced learning between both parts will be absolutely critical. We are very fortunate to work with, as you know, some of the largest companies in the world, and many of them are obviously also active in that space. We do have a successful history in the trust and safety business, as you know. That's content moderation, data labeling to some extent. And these clients approached us and say, "Can you help us also in this area?"

So this is interesting. You see some companies just focus on that one. But when it comes to how do you have access to global talent to specialized know-how? Do you have the process now how to do this? Do you have the technology and the platform in place? We are actually very well positioned to do so, and we want to further invest in this area to scale it further.

This is a client that is one of the most important clients of TP, long, long-standing client. And they said, "Can you help us?" And within month, we built this team, you see we have 650. I think the number is now even higher, machine learning [ graders ] that are specialized talent who help to provide better outcomes for the AI. And we're, as I said, setting up dedicated teams to further accelerate this area and the acquisition of Agents Only, which is a crowd-sourcing platform, focuses exactly on that space, how to have better access to global client, not as a full-time employee but as a gig worker, to drive data services.

With this, I leave you now in the good hands -- before the summary of my colleagues. You will hear first about from Anish and Akash how does TP.ai allows us to capture more opportunities in the AI value chain. Secondly, we'll hear about Miranda, our Global Customer Officer; Mamta, our Global Head of BFSI; and Himadri, our Global Solution Experts, how we drive growing the core and extending the core business process services business, so the first 2 pillars of the strategy. And then we have JC, Sherri and Jeff Cordell, who is our CEO of Health Advocate, explaining how we do the same thing, growing and expanding the first 2 pillars of our strategy for our Specialized Services before we come to a wrap-up.

So have fun, enjoy, ask tough questions because that's -- we are here for you today. And with that, I hand over to Mike.

Fantastic. Clear vision on where we are today and where we're headed in the future. Anish will talk us through the [indiscernible] how does TP brings all this together into an AI platform that solves complex, interesting problems for our clients.

Anish?

Can you hear me? I'm going to move this.

Okay. So AI has moved much further than the orchestration stage at TP. It is at the foundation of how we reimagine the operating model, how we deliver outcomes for our clients and how do we drive growth. TP.ai FAB is a new fabric of agentic AI systems, outcome-driven Copilots, with customers at the center, driven by intelligent automation. It is embedded in everything that TP does now, from customer care to vertical-specific solutions like underwriting, to finance operations and to internal functions like talent management.

I'm Anish, I am Head of AI and Transformation at TP, been with TP for about 3 years, as Thomas said. Prior to TP, I was leading the customer experience or the shopper experience for Amazon across 12 of the largest marketplaces. And I'm very excited today to be part of TP's AI forays and contribute to the transformation journey of our clients. And unsurprisingly, today, I'm going to talk about our AI business and our AI FAB, which is a version of our enterprise-grade modular AI operating system.

I'm conscious in the next maybe 15 or 20 minutes. My conversation may get a bit more technical. So I'll give you a few business examples and the context as I go through the pages.

TP and our ecosystem, our industry, our clients, our peer group is at the inflection point. Customer expectations for us are rising. Cost pressures are also real. 41% of our CX customers see Gen AI as an incremental as well as transformational element of this strategy by 2026. As we look at it today, AI is good enough to rewire operations, only if you know it to how to do it at scale. And TP.ai FAB is our answer.

There are 4 distinct trends that I want to talk about today that informs and leads TP strategy. First of all, AI is at the top of every client's agenda, every client that we currently service and every client that we currently pursue. And the client could vary from tens of billions of revenue in each marketplace that they operate, in each country that they operate in, and they could be in many countries, to a client which is less than $1 billion in revenue. Every such enterprise is prioritizing AI adoption. Gen AI, which was once confined to the boardrooms or to innovation labs, has now moved from experimentation stage to the implementation stage and at a very, very fast pace.

Hundreds of our existing clients and potential clients value 3 things. First of all, AI has to be scalable. It has to be secure, and it has to be based in real-world use cases. For instance, a health care client of TP enables self-service systems, which are now beginning to handle claims and appointments with north of 90% accuracy, and creating huge amount of capacity within TP for higher value [ player ]. Why it is important for us to know is because of that released capacity, we are going upstream and downstream and creating far more value for our clients and, in turn, creating far more value for ourselves.

The future-ready workforce is currently a challenge. The challenge is determined because of the increased demand, which is determined by the hype that AI has created, there is sort of a frenzy going on. Also, sort of expectations, which are not mired in reality, but at the same time, there are some valid opportunities that our clients want to pursue.

Underlying this lack of capacity are 2 major reasons. One is the lack of real-world experience, the curated cases that are relevant to our clients, and lack of industry domain. That's what our differentiation is. We have decades of experience across major verticals, where we control the outcomes, we deliver the outcomes, and we can bring it to the clients' benefit. That's our competitive edge. We bring deep domain and transformation mindset, and coupled with the horizontal solutions that you see all around us, it is an unbeatable combination at our scale.

Important to note, the mushrooming of AI companies that you see all around us, they are right now even still more focused on the horizontal solutions. And I'll give you one more example of that later in the slide.

Secondly, clients want end-to-end scalable solution, not a point solution, ownership of customer outcomes. They have moved beyond asking for a -- asking a question that what can AI do for them, to the next level of granular question, is how can TP embed those in their enterprise operations and do it sustainably and do it with scale.

Customer expectations have changed. Thomas alluded to omnichannel. They expect seamless, personalized and omnichannel experiences and only when it is applied thoughtfully with agentic AI solutions and human experts working dynamically together and self-learning.

In conclusion, TP is not just riding the wave. There is a big wave going on, and we would love to ride it, but it is not just about riding the wave. We are reengineering from inside out. We have the deep domain capability of more than 40 years of deep domain expertise and transformation roadmap that we can bring to our clients' benefit, that's what we are doing now. It's all integrated in our TP.ai FAB and which is now ready.

Let me ask -- bring this to reality. The ask of those clients, existing as well as the potential clients, is causing to unfold the future of CX, with both domain services and agentic AI capabilities. FAB's TP orchestration platform, which is modular and AI-powered, which weaves an agentic AI experts into vertical-specific solutions. There are 4 key tenets of FAB that I would want you to register. First is fairly intuitive. It's structured in 3 modular layers. At the top is the blueprint layer. In the middle is the AI orchestration layer. And at the bottom is the foundation layer. The nature of these layers allow us the flexibility, speed and use case orientation.

This structure is designed to scale across clients, even though you might build an AI-curated case, which is contextual only to a client, but a large part of that is portable. So it is portable across clients. It's also portable across different industries, different verticals because it's a combination of different horizontal services. Eventually, you can do it in a region, you can do it in the second region. So it is portable across geographies as well. Each layer is there for a reason. It adds unique value, whether it is foundation layer or the blueprint layer, that I'm going to talk about in the next chart.

In the blueprint layer, this is where FAB for us becomes a product. These are outcome-driven use cases for our clients and TP's most compelling transformation differentiator. These are also the processes that our clients very intimately relate. TP creates the AI vision, which is embedded in our vertical knowledge, we redesign the CX, and we improve fulfillment of end-to-end processes. I'll give you a few examples on what we mean by the vertical knowledge here.

If you look at financial services, from account servicing to customer acquisition, to anti-money laundering, risk services in financial services, claims, customer onboarding, policy issuance and insurance for horizontal services as well for closing -- accurate closing and timely closing of books, filing of 10-K and 10-Qs, automation of manual general entries and finance and accounting processes, and reducing the defect, improving the shoppers' experience for some of the world's largest online retailers. Those are the use cases that we're talking about.

You can imagine the blueprint as the pre-orchestrated package, which combines AI, expert and logic together, which is tailored to a real use case. You can also visualize them as pre-configurable small blocks of small language models, or SLMs, which can either be sequentially deployed, or in parallel, in combination with human expert and with agentic AI and they include orchestration logic, data flows and domain-specific tools. That's how TP is moving from capabilities to solutions.

Simply said, if I were to explain it little bit -- a little less of technical terms for our clients, it's an end-to-end process redesigned with envisioned AI tools and human experts, which are dynamically working together and self-learning, self-serving. At the bottom of the page, there is an example where our horizontal solutions and vertical capabilities come together to create state-of-the-art banking and financial services process that Mamta and Himadri will talk about in the next session.

AI orchestration layer is the heart of FAB. Every task in the FAB starts with AI, that's the default. TP at this layer combines AI agents and human experts and lets the system decide. The system decides the following: Who should handle the task? What logic should apply? And when should we escalate? This FAB logic also incorporates outcome orchestration, task allocation, real-time learning for the machine, knowledge and the feedback loops. Even if the task and the question goes to an expert, AI doesn't disappear from the scene. It sits by the side, learns, provides live prompts, sits on the top of the knowledge base, suggests the next best action, which the agent can then communicate to the customer.

In this example on where the interaction went actually to the expert, we are talking about AI, making the expert more efficient and more effective. But at the same time, machine is learning, AI is learning, expert is making the machine learn. So therefore, AI will become better over a period of time.

This orchestration layer is not rule-based. It's dynamic. It's based on the complexity of the use case. It's based on the customer intent, and underlying this dynamic process is our probabilistic model that supports the decision tree. And the best part, as I said, it gets smarter over a period of time.

Finally, the strength of this orchestration layer lies in the modularity of it. There are proprietary tools, there are TP tools that can be selected or deselected depending on what the customer wants us to do, the amount of ownership that we have of the process. But due to the demonstration of improvement of customer outcome that the client wants to achieve, they very easily allow us to go in an accelerated way to upstream, downstream. And what that causes is even greater impact on customer outcome and a greater impact on the value share that TP derives from that engagement. This is how TP will build, scale AI that works.

Finally, the foundation layer. It is of utmost importance to have scalable, secure and globally accepted ecosystem of hyperscalers. The foundation layer is where the trust and the scale begins not just for our clients, but for TP as well. FAB is built on enterprise-grade architecture, cloud first, and LLM orchestration and model agnostic foundation.

Why is model agnostic important? There are new clever models, LLM models, that are coming up, which are also more efficient. If we are model agnostic, it gives us unique ability to be future adaptive and bring those into our fold. The core AI skills of reasoning, summarization, classification and automation are built into this layer. You might imagine and question that why is this important, because this is something our business clients may be oblivious to. But this is where the most, importantly, secure data flow resides, observability happens, prompt orchestration takes place and LLM integration happens. So foundation gives FAB its ability to scale securely.

Now let me bring the FAB chart again with all the 3 layers defined, with all its critical parts displayed. This is FAB under the hood for you. And let me try to make it real with the help of LEGO analogy. Imagine that the task for us is to build a LEGO car. And we have a team of perhaps smart robots who can do that. Imagine each LEGO brick to represent a small AI capability. One could be of language, one could be of vision, one could be of decision-making, one could be of sequence. And a car is envisioned on how will it function, what parts will come at what time and what different components needed to be added. That design phase of how the car should function using these LEGO bricks happens at the blueprint layer.

And in the past, humans had to pack and place each brick separately, carefully planning. But in the AI orchestration layer, there is a dynamic handoff happening between those robots that I alluded to and the human experts. Now between both of them, they'll pick the right bricks, and bricks here are the AI tools and APIs. They'll decide the right order, which is the sequencing or the workflow, which is embedded in our framework. They'll make the adjustments, which is learn and adapt, and they'll deliver a working car.

Now you might imagine that this is completely done by robots. It is actually not. Because robots even today in any of the use cases that you'll see around us, and for a long while, will not have the capability to do end to end. If I were to extend this, robots may not know how to plant a steering wheel, how to attach the wheel so that the car could move. That human supervision layer will continue to become more and more important. So the result is a fast, flexible, intelligent system that doesn't just reside in a predefined task.

Foundation layer is a little difficult to explain in this analogy, but I'll try anyway, because that's where it scales. If you were to extend this analogy a little bit more crudely, you would imagine the safety policy, the purpose of the car and even the highway infrastructure would become part of this analogy because it ultimately allows our analogy to scale and the car to function.

Key points I would want you to take away from this structure. Our FAB structure is built on the bedrock of the best hyperscalers that you can find in the world today: Microsoft, Amazon, Google. And it is LLM agnostic. It drives scalability, it drives security. Intelligence layer of our AI orchestration layer is built with our proprietary tools, and it's built in partnership with our AI solutions. Orchestration is owned by TP. It is modular, and it doesn't need to be bought all at once. You can buy a part of it, you can buy additional part of it later. It gives our clients unique ability to work with us and start small and eventually grow. Deep vertical domain expertise that's responsible for customer outcome design with embedding AI and the blueprint layer, that's our main differentiator. With this, we firmly believe as a unit that TP will be the industry leader, we will be modular in our solutions, we'll work with our partners, building an open ecosystem.

And the FAB's, I'm sorry, intelligence layer actually is not a clean state to rebuild because you might imagine, suddenly, we have woken up, and we're starting to build something new. It's built on what's already proven. It's built on our decades of experience across the major verticals. On the left, you see mention of a few tools: TP Interact, TP Protect, Power Steering, et cetera. Those have been existence for years. And just in 2024, we have done more than 200 implementations. However, in today's rapidly changing AI landscape, partnerships are critical. They are a strategic imperative for us. They enable faster innovation, and they allow us access to specialized services and the flexibility to scale without being locked into a single technology stack.

I have a few names here. Sanas helps us enhance communication clarity globally through real-time voice translation. And TP is the exclusive reseller of speech tech for many of our largest companies -- client companies across the world. Parloa for us brings agentic AI multilingual voice expertise that complements our already huge scale on voice-heavy interactions. And we have Malte here somewhere. He is a CEO and Co-Founder, and he would love to spend some more time later in the session with you talking through what he brings on the table and what partnership is doing for our clients.

Our collaboration with Ema accelerates Copilot deployment across enterprise operations. And we have Surojit Chatterjee here, and he'd love to talk about his case studies when we get to that in a minute. We are extremely proud and excited what this partnership ecosystem can do for us. And what you see on the slide are just a few. We have already earmarked EUR 100 million for investing in those partnerships.

Finally, not stealing Akash's thunder, he's going to talk about that Agents Only, we are very excited and proud to present Agents Only, which is a flexible and gig-enabled workforce model that can be intelligently matched to client's demand using AI. And we are only unfolding the monumental opportunity that presents on this.

Together, these partners that are already in our fold, and many more that are going to come in the next few months, these partners extend our capabilities far beyond than what a single platform can deliver for us. They allow us to remain modular, adaptive and fast-moving. When we focus our internal investments to what differentiates TP the most in the blueprint layer and in the orchestration layer. This allows TP also to become more resilient, more scalable and better positioned to win in the AI-powered customer experience economy.

Finally, FAB capabilities across these 3 layers enable us to go faster and deeper into new markets. That's something that Thomas spoke about in his presentation earlier. As these 3 layers, and AI in general, reshape how the industry is going to change, reshape and how we're going to strategically expand into our high-growth adjacencies, digital marketing, consulting, technology, data services, that together represent north of $600 billion total addressable market, those are the markets that are present to us as new, new markets, and those are strategically significant opportunity for us.

What sets TP apart is that we are not entering these markets as newcomers. We are entering into this market with FAB as the foundation. And our real execution expertise and transformation playbooks at scale, we are bringing that to our clients' benefit. In doing so, we move up the value chain owning not just the execution, but the end-to-end customer outcome, therefore, expanding the value share that we get from them.

I'll give you a couple of points. For example, in digital marketing, we are using customer, consumer intelligence enabled by AI to drive engagement, build brand equity and power growth for us. In consulting, we are using future-ready transformation models, which are grounded in AI feasibility, and we propose what's possible for our clients. In many of these instances, our clients trust us to actually execute these transformation roadmaps. And whenever we do, we transform those client operations, obviously, impact their customer experience, outcome, reduce cost, bring scale to them. But in turn, we do far more business with them, expanding our value share.

In technology, we deliver Gen AI services, Gen AI-infused IT services and cyber-resilient infrastructure, which is built for scale and speed. Finally, in data services, we enable everything from AI model training, descriptive, diagnostic, predictive and prescriptive analytics that soon Akash will talk about.

Thank you. Am I audible? Yes.

Hi, everyone. Good morning. My name is Akash. I've been with TP for about 5 years, in the industry for about 25 years. Very fortunate to have built the trust and safety practice here at TP, which is a little over EUR 1 billion now. And very excited to talk about the new journey that we are taking in the data services space.

So data services, when we think about AI, there are 3 things that are required in AI: energy, chips and data. And data is actually the foundational fuel of AI. If you think about any large language model, if the large language model does not have curated, high-quality, diverse data, the outcomes that you will get will not be appropriate for what you're looking.

From a TP standpoint, we are very excited to share this new strategy with you. From a data services standpoint, TP has the end-to-end data services available for our clients. What does that mean? I think when we talk about data, a lot of us understand that data is nothing but data annotation and labeling. While that is one of the major parts of data, we also have data analytics, which is huge, as well as data engineering/data management. So we serve end-to-end all our clients.

What clients need here is a provider which can provide high-quality, modular, scalable, secure data for their enterprise. All the companies, beyond technology companies, will soon look at service providers like us who have the understanding and experience in delivering data services.

What's happening in the industry? So if you think about the last 1 year or 1.5 years since the time ChatGPT has come into play, a lot of us, our searches have changed. We no longer now generally go to Google as the first place to search, but then we do ChatGPT. And if you look at the last 1 year, there have been 48 billion searches on ChatGPT alone. It's growing at 80% plus. While we say that we don't Google it, Google also has their ChatGPT version, which is Gemini, which also has been enabled within searches. So that's happening.

Why is that important? Again, everything that comes out of the search is data. We are consuming data. So that is where the entire piece comes together. The other thing is that the AI ecosystem is expanding. So we are not only talking about industries, which is tech, but across different industries. When we look at asking questions from ChatGPT, we are asking various questions, be it an academic question, be it a travel question and so on. So it is going through different industry segments, it's going through different languages and modalities.

Data is the foundational fuel. This is actually driving not only just large language models. But think about autonomous vehicles, think about robotics, all of that requires high-quality curated data and secure data. And that's why TP is well positioned to deliver because of our great strength in terms of our vertical expertise across the different domains, great expertise across getting the different languages and the context awareness that is required in data services.

Now when I talk about TP.ai data services, it's an end-to-end platform that is available for our clients. What I mean by that is that it does not only spans across horizontally across all the industry segments, but it also looks at the entire vertical AI chain -- AI spectrum. We don't only look at consulting for AI implementation on data for our clients, but we go through the full end-to-end cycle where we build operationalized, as well as govern it, right? We also look at the different modalities in terms of text, images, audio, video, et cetera. That's why clients prefer us because they get scalable, on-demand, curated, high-quality data. They get it with the highly specialized skill resources that we are able to bring at absolutely agile and adaptable way. That is where everything comes together.

When we look at the different areas, which are growing, if you think about large language models, if you think about computer vision, if you think about physical AI, which has been the new thing that has come in, that is where data is required. Physical AI, just so you guys know, is nothing but where you have drones, you have robotics, you have autonomous vehicles coming together and gathering information. All of that information is data. So that is where this is happening.

How we help clients and why clients value us. Obviously, we are able to provide data as a service, not just curated data, but high-quality, scalable data for them. We are able to provide human-in-the-loop scenario. This is -- I think there's another thing that we always think about is that this is a onetime exercise. It is not. It is an ongoing exercise that will continue because the models need fine-tuning and the model needs to adapt to the changes in the industry that's happening. So human-in-the-loop is extremely important.

The other piece that is important that we bring together to clients is our trust and safety expertise, which means bias mitigation, which means hallucination checks, which means data -- which means the AI not generating its own set of facts, which are not correct. So that also is something that we are able to bring together.

The last thing that we are able to bring together is the understanding of all the laws and the regulatory environment and the regulatory framework that's coming across the different countries. Again, this helps us because we are across 100 different countries and our understanding of the country and the regulation.

The other piece that they love about TP is the flexible, highly-specialized skill resources at scale that we are able to provide. Now let me spend 30 seconds on this. When I talk about highly-specialized skilled resources, what are these? These are mathematicians, physicians, doctors, lawyers, opticians, pharmacists, radiologists and so on and so forth, right?

When we talk to ChatGPT, we ask any question -- and ChatGPT is just an example, there are many more large language models. We ask any question that may be there, right? We can ask a medical question. We can ask a mathematics question. And it has to provide an answer. The point being, when you look at the answer, those answers needs to be accurately looked at by someone as a human who's an expert in that space. We are able to provide that at global reach, at scale and at a cost that is really attractive to our clients. So that is why TP is at the right place for our clients.

Now this is not new, okay? So I want to emphasize on that. This has been going on for the last year or so, where we are growing this practice significantly. Thomas touched upon this earlier. For one of our largest global tech clients, we started with data annotation and labeling, and it has grown tenfold, right? We have over 800 to about 900 people doing this work. We've moved from doing onetime data labeling exercises to actually changing the brief over a period of time where we are looking at LLM -- we are looking at LLM summarization, we are looking at prompt quality and so on and so forth. So we are going up the value chain as well in terms of managing this work. So this shows that we don't only understand what the client wants, we are able to provide it at speed and agility.

Before I move on to my last slide, I also want to mention the way this is getting delivered across the AI spectrum. So there are 3 ways that the AI companies are reaching out. One, through managed services providers like us. The second way is looking at crowd-sourcing providers. And then the third is pure-play platforms.

What we are doing is we have a three-pronged strategy. Obviously, we are really good at providing managed services at scale at a cost where the client wants. We are also partnering with platforms where we can be the managed services provider and help them in their AI journey. But we are also moving into the crowdsourcing space, where we are able to find these specialized skill resources who do not work full time, but they are able to provide their expertise through a gig space. And this is why I'm really proud to announce the Agents Only acquisition here today. Agents Only is an AI-enabled crowdsourcing platform that connects global brands with curated on-demand workforce, highly-skilled workforce. We are able to provide it with speed and the quality that is required. We are able to provide it at a scale that our clients expect. And we are able to provide it at the flexibility and the security that they require -- that they need.

So in summary, while AI is really changing and disrupting industries, TP is well positioned because of our global scale, our human-in-the-loop network, our highly-specialized skill resources, platforms like TP.ai FAB and Agents Only. We are not just looking at AI as a partner here. We are actually creating the infrastructure layer for AI.

So very excited to -- for you to experience data services with us. And handing it over back to Anish.

So I think the first question that you'll wonder about is: Why now and why are we supposed to win in this market? That's underpinned by 3 core messages that I want to talk about.

First, TP comes from a core practitioner DNA. Unlike a lot of advisory firms that you see around, we build and deliver where AI meets the real world. Our solutions are grounded, agile and also impact-driven. We have done 1,500 installations, so to say, across 700 clients already.

Secondly, we have scale-ready talent, which is north of 3,000 specialists. Those specialists have solution architects, software engineers, data experts who are already working through all our existing clients as well as a huge pipeline of opportunity that we see. We have over 10,000 supervisors who are getting trained on AI and transformation as we speak. That will increase our reach to fourfold to where we stand today.

And finally, we have operationalized TP.ai FAB. That's our backbone, that integrates data, models and Copilots across functions. It's a scalable way to embed AI into every transformation play that we make. And these aren't just future ambitions for us. It's happening. It has happened in our context already.

We have built a social media hub for a beverage player resulting into 900,000 new followers, 850 million new impressions. For, a global airline, 15% productivity, for a U.K. logistics player where we built a digital engagement platforms, and so on and so forth. Akash has already spoken about the digital case. This is what makes TP credible, that we do not just advise on transformation, we deliver it, and we power it by AI.

I will now -- I'm going to elaborate now on these 3 to give you a little bit more context on what actually it means beyond just that specific case and why is that case important.

First, on the technology side. TP partnered with a U.K. logistics carrier to deliver a full-scale AI-led transformation. That resulted into 15% increase in agent productivity and 2 million productivity gains. We used TP.ai FAB as the backbone, deploying Gen AI power interaction analytics for 100% of the customer queries with the real-time agent assist and next best action models. The solution also included Genesys CCaaS integration, AI chatbots for improved self-service and accent transformation technologies.

This is a full suite of FAB that I earlier spoke about at display from blueprinting layer to the foundational layer. And therefore, why is this important? Why is this single case important? Because it demonstrates TP's ability to combine deep domain expertise with advanced AI and tech, and we deploy it with unmatched velocity.

Earlier in the day, I was speaking to one of the investors, and we're speaking about the need of being agile in today's world, because solutions will come. The differentiator lies is how can you quickly scale it up faster than anybody else? And how can you determine the early bets that will work? At the same time, early -- detect early will not work, so you're not spending a lot of energies chasing the dreams that will not happen. And we have demonstrated these applications in complex, really complex service environments.

The second case is about our transformation play for a leading global airline for consolidation of its contact center operations. That delivered north of $10 million of annualized savings. We leveraged our deep domain expertise on the airline domain knowledge with CX design and AI capabilities, conducting demand and supply analysis, deploying AI simulation models across, completely redesigning the target operating model, resulting into a 50% reduction on the onshore centers that they have and almost 60% of the workforce moved into a remote model. And this case underscores our ability to lead strategic AI-informed consulting assignments that can optimize cost, elevate experience and drive sustainable value for our clients and for TP.

In this final case on digital marketing, we partnered with a global consumer products company, one of the largest in the world, to design, build and operate a centralized social media hub. It transformed how the brand interacts with their customers across the platforms, an omnichannel communication engine that enables real-time AI informed engagement across social, chat and digital touch points. This drove exponential growth in digital reach. Their impressions increased from 1 million to 2.5 million, resulting into a disproportionate increase on revenue on the TP side as well. This program demonstrates our ability to deliver marketing transformation, blending consumer intelligence, experience design and engagement to drive growth, brand equity and loyalty.

My 3 closing comments before I hand back to Mike. TP.ai FAB is an industrial engine that's designed to scale. It brings the best of TP's domain knowledge and cutting-edge horizontal solutions. Secondly, we will accelerate the deployment and be the industry leaders with best-of-the-breed tech with help of hyperscalers. And finally, this allows us to expand our total addressable market dramatically. It allows us to get into margin-accretive accelerated growth. And this will become a self-sustaining flywheel for TP. Thank you.

There's a lot of rich content there. And just a couple of points that strike me as I think about what we heard. One, all of our clients are talking about AI. It's at the top of their agenda. Fortunately, they're all talking to TP about how we can help them through that journey. TP FAB collected all of our best practice knowledge, our deep domain expertise to allow us to address those needs in a systemic, repeatable way that doesn't require reengineering each time.

As we generate value for our clients, they offer us the opportunity to solve new problems with them, opening new lines of business, growing new areas of work. And as Akash mentioned, this opens completely new markets in things like annotation and labeling and actually helping build AI systems for some of the world's largest companies. Truly an incredible opportunity that we face into and one that gives us all a lot of energy.

Up next, we'll take a look at our BPS, our core business process services, evolving that core with these new tools capabilities. Last, we'll cover the vertical expertise and how we're looking at the entire value chain of a particular industry, where we can play, where we can help solve those problems.

So I will also introduce Miranda, our Global Chief Client Officer, to the stage. Thank you, Miranda.

Good morning. Thanks for being here today. It's nice to see some familiar faces. As Mike said, I'm Miranda Collard, I am the Global Chief Client Officer. And I too have a very distinguished career. I've been at TP for 31 years now, starting as a TP frontline expert. And ever since, I've been passionately serving both of our clients and our employees. So happy to be here.

So I have the distinct pleasure of covering for everybody today a couple of pillars that Thomas mentioned earlier, which is our global business processes, the core growth with AI as well as the verticalization.

So first and foremost, I have great news. The market is growing again. So this is very, very good for TP. We have significant opportunities across the board, and I'll start first with some of the key market trends that we're seeing.

First and foremost is the AI and the humans. Really what is the top of the agenda for all of our client partners and prospects is that organizations are certainly embracing the hybrid model, which really I like to define as an enablement in the terms of we need solutions to bring forward that help the frontline with a greater level of emotional intelligence to really meet the customers where they need to be met and that our partners get the outcomes that they're looking for.

Second is the partner consolidation, and any near onshore delivery opportunities are certainly accelerating as they do in these times. These are not new objectives inside of our business. They're just being met differently with the wave of AI and all the rich architecture you heard from my colleagues, Anish and Akash.

And then last but not least, AI is driving demand for that integrated customer experience platforms. So we need to meet with the technology and the solutioning that comes forward and that allows a roadmap for that enablement that I explained before.

So accelerating the growth. First, the first pillar, and Thomas shared, is to grow the core enabled by AI. How we're doing that? We're going to share a couple of examples as we move forward of what that looks like from our rich and deep expertise. But first, our strategy, of course, is to use that tech. So it's our responsibility within the partner network to bring forward all of this incredible capability and solutioning that you heard today. You can see that it's deep and rich, and for people like me inside the organization, it's exciting, because this is where the growth comes from.

There's an example of this, of course, in the banking, financial, insurance -- services and insurance group that you'll hear from Mamta here in a bit. So embedding that AI across all of our core services. We enable with the TP.ai FAB, we accelerate the transition of in-house and capitalize on that adoption. This is why you'll hear also that the verticalization is so critical for us.

And then we have a multitude of verticals. And by the way, I could talk to you about them all day long. But we're highlighting BFSI here. But just to give you a small example of another vertical with the retail and e-commerce, since we're all consumers here, we can all relate. To personalize that online journey across the retail and the e-commerce space, it reduces the churn that all of us have experienced as consumers and in the business, and it really optimizes those business outcomes. So you think about the retail space and the retention of the customer, bringing the customer back, having the repeat purchases and all of those are extremely important for our client base.

And then we'll get into some -- the vertical plans in the extended space. This is one of the most exciting things that we get to do in the solutioning group here at TP. We can now expand inside of those verticalization that -- and you'll hear some examples of that, we build those industry-specific end-to-end AI-enabled offerings based off the very deep expertise that we have in our leadership teams.

In BFSI, and I think everybody probably has something to say about BFSI in this space, but if you think about the amount of workflows, the amount of domain experience that you really have to have a knowledge of to transform, this is a key critical component for us. Having a full understanding of the end-to-end, the front-end customer journey, the back-end processing and all of the thousands of workflows in between is a significant opportunity for TP.

And from the retail e-commerce space, we can codevelop now those solutions. So if you think about demand forecasting, when are those interactions going to come in? When can we automate? How can we capitalize on it? Because as you think about the mundane automation of repetitive workflows, what's going to happen and needs to continue to happen is when you need -- when the customer needs to speak to you, they need to speak to you. So demand forecasting is incredible. And then, of course, like inventory optimization, automated returns processing, all of those things that allow us to take friction out so that now, with our emotional intelligence training and program, you can really show up in those ways that you need to and the customer needs to hear from you.

So we are going to -- I know that everybody is asked to speak to our clients over the years. So we have invited 3 of our esteemed partners here to join us in a panel conversation that will be moderated by the wonderful Mamta Rodrigues. So I'd like to invite you up, my dear friends. And I'll allow them to introduce themselves.

Okay. Hi, everyone. Thanks, Miranda, for that amazing introduction. I'm Mamta Rodrigues, and I run banking, financial services and insurance. And I have the easiest or the hardest job today because, you guys are going to have to help me. Keeping these 4 quiet about this space is going to be interesting. So I'm also going to play time cop a bit. We've got 20 minutes. We're going to start with introductions, and we're just going to start here, and we'll go that way, Paul.

Sounds great. Good morning. Happy to be here today. I'm Paul Vincent. I am a senior adviser with Boston Consulting Group. I have been doing that job for about 45 days, so I know it all. I'm on panels, so I'm excited. Also have 25 years of banking experience before that, both at Capital One, an information-based strategy company in the financial services industry, and USAA, the leader in customer experience. And roles that have spanned operations, channels, product, strategy, transformation and eventually being President of the Bank at USAA.

Good morning. I'm Lance Gruner. Also very, very happy to be here. Thank you very much. I've been in the industry well over 30 years and most recently have been with Mastercard as the Executive Vice President of Global Customer Care, responsible for service and support in 210 countries and 74 different languages.

I'm Lisa Stoner, I'm currently the Vice President of Support Operations for Pinterest. Previously was a buyer of services at TP, and I was the Vice President of Products and Support Operations at Meta, as the Vice President of Global Support at Meta and as the Head of Global Support for Uber.

Okay. So let's jump in. The first question, Paul, I'm going to start with you. The environment for AI has changed radically. With your multitude of years of experience, you've seen a lot of the evolution of, not only AI, but a lot of technology that's come underway. As you look at AI and disruption in operations, what do you see the change in customer expectations, in the omnichannel and multiple channels? And what do you see as the change now? And then what do you see as foresee coming forward?

I don't know if I like starting with the multiple decades of experience. That will age pretty quick. It's interesting. We used the word unprecedented so often right now, I think it's losing its actual meaning. But we are in a significant disruption from a cycle perspective. And it starts with the consumer. And so if you think through the average consumer for financial services, you look at digital commerce, which has doubled in the last 5 years, you look at embedded finance, your ability to pay while you're in that transaction with debit or credit or alternative forms of payment, that's up 5x in that same period of time.

And we're just getting to things like agentic commerce, so where you'll have an agent that is actually shopping for you. If you watch PayPal or Visa right now, they're introducing agents for the first time. And you'll give them your parameters, what you want to look for, the value you want to pay for it, and it will bring back the best options for you. So you've got a wave of change coming from the consumer, which ultimately means we have to change as well as financial services companies. And we are looking at AI and how it plays a role in the entire value chain, ranging from targeted marketing, underwriting, servicing, collections, you name it. There are roles that will play to make that experience more efficient, more effective.

And by the way, it's not a one-size-fits-all. You've got a digitally native group that has now grown up with the phone, the younger generation. But you also have those moments that matter, those highly emotive experiences that Daniel talked about, where they'd still want to talk to a person, want to go through that, want to know that there's someone they trust on the other side that's there to help them. And so we're dealing with needing to handle the broad needs of different groups in a vastly changing environment.

If I could add, I think not only that, but adding on you've got the speed of change that's occurring today. You have post-pandemic, customers' expectations have significantly changed. You have the need of customers -- or companies to anticipate where the customers are, the need of the companies to provide services has changed. That change is now becoming so -- the speed at which it changes is the need to use AI is ever more important, the ability to have more data at your fingertips, the ability to anticipate customers' needs and physically what that change that is necessary to compete in today's market.

Anything you want to add, Lisa or Miranda? Are you guys...

No, good.

Okay. What I heard from you guys...

That's a first for Miranda.

I was told to be...

What I heard from you guys though, which is paramount to also what we heard from Akash earlier, is the power of data. And we're just getting started. While in the space of payments, one can say, we've had data and the data lake for a while, but the monumental pivot now with AI and what I heard from both of you and the leapfrog effect that we're going to see, especially when you double down on that customer leading the way, the decisioning coming from the customer, is phenomenal.

So second question, and I'm going to start with Lisa on this one. Lisa, you've worked with TP on a multitude of opportunities that you've been at in your years. Where do you see -- where have you seen partners like TP, not only TP, working and partnering with you, especially as it relates to integration and ability to be that leader in operations and with AI?

Yes, it's interesting. So I actually started as a competitor to TP. So I worked in this space my whole entire life like many people here. So my first time buying from TP was when I went to Uber and it was pre-IPO during a time of just incredible scale, right? And so we were in the infancy of the AI revolution then too. It just wasn't sort of as top of mind as it is now.

But at the time, TP was just a leader in just enormous change for us. So think about pre-IPO Uber, think about post-IPO, think about the COVID years, the need to vastly expand in Eats business, right? And so TP was just incredibly well positioned based upon the global footprint, nimbleness, the ability, the huge range of services, the ability to provide things like local languages in India, like all the way to highly advanced enterprise support.

So fast forward to Meta, and I had the privilege of working there during the year of efficiency. And think about, again, like a time of a tech-enabled company going through a vast change. And again, Teleperformance was a leader.

Today I'm at Pinterest. Again, significant change. And just to contextualize Pinterest for people. It's a company that where AI is a core competency. So it's at the heart of our curation signal. So as pinners, people using the platform, are, as an example, searching for outfits or looking for things to buy in the e-commerce world, the curation signal is key for Pinterest.

AI is already making advertisers more effective and more performant. That's a huge change in the ad space. It's making our employees more efficient. So to put a number against it, our CEO announced in the last earnings call that 25% of code was written by AI. That's an enormous change. And we're a platform that has inclusive AI, and there's an enormous demand for that. So think about being able to search by body type or skin color or hair texture. And so we found that search is enabled at 75% faster with inclusive search.

So it's real and can be quantified. And TP is an important partner for us, ranging anywhere from the trust and safety space to working with Pinners, to working again with a wide range of SMB to enterprise advertisers. So again, it's a -- TP is an important partner and a critical part of my personal and the company's strategy.

Thanks, Lisa. Anything you want to add...

I'll just double down on something Lisa said, AI is a core competency. And you've heard a good bit of that today. I think when you look at AI right now, you've got to make sure it's in the tone at the top, in the strategy. You've got early wins and use cases. You heard a lot of that today. But it really working means it's embedded throughout your organization. And so those that embed this and get your frontline practitioners that those that really know the work, fluent in AI, they're going to be the winners.

And I have to say everything that you've heard this morning about TP, the strategy is in play in reality because over the last several years, when you look at how as a company, we've had to change because our products have become complex, the expectations of our customers have changed, the ability of partnering with a company such as TP being able to go along the way and being there for us with the scale, with the information has been vital.

I just think the range of capabilities as well, right? And so both human and technical capabilities. So our needs are changing. And I'll just give a very specific example in the support space, right? So previously, you had -- I think RPA was mentioned, you previously had very different -- various like IVR is a million years ago, or declarative chatbots. And now we're in the age of Gen AI chatbots, right? And so they can take on the scripted work. We are using that as an opportunity to invest in higher-value work. And so our need is for people with greater skill, more language capability, more personalized interactions, again, like indexing toward enterprise and the monetization level of the platform. So TP is well positioned in that space to be helpful.

Yes, that's probably my favorite thing to talk about with clients, not just all of us here, but just generally speaking. Probably just between this panel, we've all evaluated tens of thousands of interactions across the space that you're in, previous roles, obviously, I have across all of the group. And one thing that always remains relevant for me in those interactions is that -- and because I've been a TP expert, I've sat at that front line. I've taken the call, I've made the call, I've done all of those things, is that we are in such a ripe opportunity to have that enablement factor for the front line. It's really unseen friction that behind the scenes just for the training, just for the nesting, just in the preparation to handle that interaction, it's monumental now what we can do for that enablement.

And like Lisa said, what that -- what happens in so doing is now you have a rich need for a high-value interaction and experience happening that requires a level of emotional intelligence that we have yet to see. But it has been the constant evaluation of what we've seen now for decades. As automation has come in, chat has come in or other omnichannels, is that this has been something that has been transforming over the decades and now it's just on its next wave of transformation.

And what happens is that this creates a rich experience for consumers and also for the employee, for the next person that's coming in and showing up, we can enable them in a way that we've never been able to before to be successful and have exactly what they need at that experience level.

And I think that's what you guys also highlighted was just phenomenal, to what Miranda just wrapped up with, and they were talking about this at breakfast so passionately, is you're investing in the talent. You're investing in the career. And the number of times I've been in front of clients in a good way, and we ask the question first, how does this help the expert? How do we make an impact to the economy that we are contributing to? And how do we make an impact to the customer experience? At the end of the day, because people matter.

So with the last question, I'm going to start with you, Lance. And we touched upon it a little bit, but let's bring it back. There's a lot. There are a lot of partners. There's a lot of disruption in the industry in a good way. How do you see partners like TP working with you as you go down the journey of leveraging AI in your operations? And how do you see that partnership working generally as we move forward?

So what you heard a lot about the strategy as a company is something as a partner I look for. And it's something that is really, really important because not only is it just about putting somebody or a head in a seat, it really is about partnering and understanding my business, providing me the data that I don't know today, allowing me the functionality to learn more about my customer.

And being closer to TP, with a company that has that strategy that says that it isn't just about answering the phone, it really is about getting a deeper understanding of your business, what your customers want and need, and to be able to solve that need, that today is the secret sauce for companies. It's really is, how do I partner? How do we use the data? And how do I use information and technology to better arm everyone involved?

I'll just add, and you said this. It wasn't just the need to get to know me as a customer of Teleperformance, but it was my customer at the end. And so coming from an organization with a tremendous military affinity, that went as far as hiring veterans, it went as far as hiring their spouses, training and onboarding military acumen as you brought new hires into the association. And just those are the things I was doing and I was looking for in the people that would partner with my organization, and I think one of the things that stood out.

Yes, I think the other thing, I'll say that I was a support specialist, right, I was an expert in the area of support, but a generalist being on the outsourcing side. When I moved to the client side, I really needed to find partners who are expert in the industries where it worked, right? So at first, it was ride share, which was category creation. But moving on to both Meta and Pinterest are platforms that are monetized through advertising. And so TP's got deep vertical expertise, and I think that's a differentiator, and especially when you're thinking about the global scale of the kind of tech companies, where I've made my most recent career.

Great. That wraps up our panel. Any lasting thoughts or comments? And then otherwise, we're going to hand it over to Himadri, who's next.

No, thanks for having us.

Thank you.

Thank you.

That was great. Wonderful. Thank you, Miranda. Good morning, everyone. Am I audible? Thank you.

My name is Himadri Sarkar. I'm very grateful and fortunate to be playing the role of a Chief Solutions Officer in an era where truly EI is getting power-enabled by artificial intelligence, as Daniel mentioned in his opening speech today. And the way we are doing this is those 3 strategic objectives that Thomas included in the section. What I'm going to do over the next few minutes is double-down on 2 of those 3 strategic objectives. Now if you think about accelerating the core with AI solutions and then aligning those AI solutions to vertical value streams, from a solutioning perspective, I feel that's extremely important for 3 key reasons.

The first, we need to protect our existing portfolio of customers. We need to grow the count of new customers. But additionally, we also need to improve the revenue per customer. And for this discussion, what I've done is I have shortlisted 4 themes, 2 for each, of the strategic objectives that I'm going to talk about. What we have realized over the past few years working with clients is that, early enough in the process, when we are building the opportunity roadmap and creating a very clear, articulate business case, we would need explicit buy-in from our clients. They need to be a part of the process. And the way we do this is through a collaborative design thinking process, which I'm going to talk about in a greater detail as I move forward.

But that's just first part of the story. Just creating a business case and an opportunity roadmap is not really going to cut it, unless there is also adoption of those AI solutions by our clients. And the way we do that is we embed and assemble those AI solutions in our clients tech ecosystem. I'll give you a few examples on that too.

If you were thinking that's it, that's not it. Since after that, unless now we have taken those AI solutions and anchored it to our clients' industry and vertical value chains, it's going to be very generic. It is not going to be domain specific. And that's, quite frankly, just not important for our clients. It's important for us too, because that improves our speed to market. But this entire AI and EI transformation wave, as you would agree, that's never one and done. It needs to go through a continuous, iterative life cycle, again, driving symbiotic outcomes for both our clients and us.

So let's get started. With the first one, as I said, I double-click. So how do we do collaborative design thinking? We start off with the customer journey mapping. While we're doing the customer journey mapping, we also do a very detailed due diligence on the trifecta of people, process and technology. But in order to prioritize those opportunities, we would need to obviously add some quants. And those quants really come in from those metrics or KPIs that really matter to our clients. And we supplement it with best-in-class benchmarks so that we put our clients always ahead in the game.

When we have seen this opportunity roadmap evolving over the years, we have also realized that analytics plays a very, very key role in terms of unfolding these opportunities across various segments of the operating framework. I take a few examples of products and solutions that we have worked on. Let's take workforce management.

For workforce management, we have our AI-enabled solution that works across the entire spectrum of forecasting, staffing, scheduling and real-time adherence. For training, we have a solution, AI coach, you would see a demo later on by Justine, where it's a simulated training environment for our TP experts where there is -- where AI impersonates the customer.

Let's talk about quality assurance, the perennial problem in our industry. If you think about the way quality assurance was being historically done, there were 3 key challenges. The first, very low sampling, usually 3% to 5%. Even that low to 3%, 5% sampling, just because it has to be done by human auditors, there used to be a lag of, let's say, easily 5 to 7 days. When there is a lag, it keeps widening the gap in terms of the error propensity. And even after doing all of this, the kind of insights that you get are so shallow because it just talks about where the experts went wrong. We have actually solved for that with a GenAI interaction analytics platform, where now we are doing 100% sampling, real time. And it is just not about people insights, but a well-rounded perspective on people, process, technology and policies.

Now let's also talk about what we have done as an advent of analytics as far as agents or solutions are concerned. I think one of the common voice or the common things that you would probably resonate across the discussions that have happened thus far, is that it's not AI coming in first. It's really emotional intelligence that is getting powered by AI. And as we all agree, the interactions are getting more and more complex, across all industries: financial services, health care, retail. If the interactions are getting more complex, we will probably have to have sentiment analysis as a part of the way we drive interaction management in general, be it voice, be it digital channels.

And in order for sentiment analysis to be done right and next best actions to be driven right when the agents are taking those interactions, we have done something which quite humbly, I'll still say is remarkable. We are able to do voice-to-text transcription, text-to-text transcription. While we do transcription, we are also doing reduction and anonymization, intent identification, classification and summarization, all at a record speed of 5,000 milliseconds. That is giving our TP experts to truly concentrate on driving resolution with the highest levels of empathy without really worrying or getting bothered about the mechanics of the interaction. That is what we are trying to enable with artificial intelligence.

Now whatever solutions and products I gave you examples for is that middle layer, the AR orchestration layer. And this is a diagram that Anish introduced in his segment. This AI orchestration layer, as you see, is well nested within the vertical solution blueprinting layer, which includes the customer journeys as well as the interaction channels or the engagement layer. And then you have the foundation layer where you have all the back-end apps. Now you must be wondering that there is one layer that this guy is still not talking about, which is the fact that when all of these layers are orchestrated in the right fashion, it drives a very, very differentiated and rich level of analytics, which goes far beyond just CX and productivity metrics.

Now you're able to understand customer lifetime value proactively. Now you're able to understand what is the propensity of customers churning and just not one number by different persona and customer segments. All of that comes to life. What I'm going to do as I deep dive into the vertical plays would be to double-click on 2 parts of this slide. The vertical blueprinting layer, and the way we are driving this as a -- not in a one-and-done approach, but a continuous iterative life cycle, which is really the layer which says operational impact guided by analytics.

Now whenever I show this slide to somebody, they say, hey, come on, I've seen this. How is it different? I'll tell you how is it different. Most of the times, a key factor that aligns customer journeys with internal processes is ignored and which is that if customer journeys aren't linear, shouldn't processes be nonlinear too. We have addressed this. When you see this map, you will appreciate the way we have aligned all the AI solutions to a blend of customer journeys as well as taken away the silos that usually exists into front, middle and back office by creating a more one-office integrated solution that puts everything together.

By virtue of doing this, we have been able to create this kind of a view for all industries. We have been able to also align our homegrown proprietary solutions as well as our partner-enabled solutions such as Ema, Parloa, Sanas onto this framework. That gives us the added energy, agility and acceleration to go to the market and probably pick up any use case, which our clients might be struggling with as of now and drive the right level of efficiencies and effectiveness.

But like I said, this is never one and done. So it has to be a continuous iterative life cycle that aligns all of these AI interventions to not one metric, but to a hierarchy of metrics so that we keep measuring, we keep codes correcting, and we also save a lot of dollars on implementation costs just because it's no more a one-to-one relationship, it's a one-to-many relationship or in some cases, a many-to-many relationship. What our clients are appreciating the most about us is they're saying, hey, you know what, if earlier, we were running 10 projects to solve for something, now we probably have to just 2 projects, thanks to the way you have helped us orchestrate the analytics in terms of driving KPI measurement.

I'll say one more thing, which is why this one-and-done approach is needed. What is so dynamic about it? In my view, there are 3 things that are very dynamic. The first part is our clients' products and service landscape is constantly evolving. They are not settling to status quo. That's one. Second, our clients' consumers or end users, their needs, behaviors, preferences are evolving. A lot of those consumers might be in this room, us. And last but not the least, AI in itself is on a dynamic path.

So in my conclusion, this is what I would say with a hell of optimism. I feel the way this approach is all bringing to life with TP.ai FAB, we will be able to transform confusion to clarity, but more importantly, enable our clients with a very clear call of action to driving tangible outcomes and which is going to be symbiotic for both our clients and for us for those 3 things that I mentioned initially when I started off, which is helps us protect our existing portfolio of accounts, helps us grow new customers as well as improve the revenue for each customer.

Thank you so much.

Can you hear me? Perfect. I've already introduced myself that I run banking, financial services and insurance. I'm going to take 30 seconds and also share with you my background and why I chose TP as my home for the last 5 years. Prior to here, I have 30 years in total of experience in banking, financial services and insurance in companies such as American Express, Mastercard, Visa and Synchrony Financial. And one of you asked me a very good question out in breakfast, which was what did you do in those 25 years prior to TP? And I ran operations, product, marketing, servicing and lastly, ran digital transformation at the bank that I just came from.

I have the best job. And what keeps me excited about my job is that I spend every day thinking about my clients. I spend every day thinking about the old me in what we do today. So with that, let me get started and share with you the key highlights of what I'm going to cover today. Miranda alluded to this earlier. Thomas spoke to it earlier. We are growing again. This is an exciting industry. And when you look at my vertical, which is what I'm going to peel the onion on, there's even 50% of growth that's existing in the captives, an opportunity to start outsourcing that today hasn't happened, and we'll get into that.

Two, AI is creating opportunities. We heard it from our clients up here as well, creating opportunities for our talent, creating opportunities for ourselves, making educated decisions. So when we, all of us, you're all customers in BFSI because you all have a card and wallet, you all have a bank account. We all have that. And AI is creating those opportunities to create special moments and memories in those engagements that create growth, that create expansion in those lines of business and ultimately create stickiness, not only with our clients, but the consumers' last mile as well.

We are uniquely positioned to win in BFSI, one, because 65,000 experts in this space globally wake up every morning thinking about the customer, thinking about what keeps them up at night and solutioning to what you saw with some of the examples from Himadri on how we solution for the customer, starting with the persona, building down to the customer journey and then proactively many times, sharing with our clients as well. That also results in trust, trust that you hopefully saw on stage, but trust for us to be the leading partner when they have a question or a conundrum, we're the first that they start for.

None of this is possible without the foundation layer that [ IC ] TP.ai FAB built playing and has been playing for us and is part of our DNA for years. Now we're rapidly scaling even more so on that as we verticalize and invest even more from what you saw from Anish and Akash's presentation.

So I'll jump in. Growth, I already covered on the right. And what you see on the -- on your left actually is the industry trends. There's a multitude of trends, and you know this space just as much as I do for BFSI, but I'm going to cover 3 things. AI is top of mind and top of the agenda for my customers. An example of that is in the insurance space. In insurance, there is 76% of U.S. insurers, a phenomenal number that are leveraging AI first in their strategies. They're building out their integration layers and the next generation of transformation with AI first.

They're starting to lead the way as well, whereas in the past, many times, there were other verticals and BFSI is leading, especially in insurance on that front. Outsourcing is delivering efficiency and efficiency gains. That's been there. That's table stakes. The reality of the pivot also is that there's loosening in regulation, and that backdrop is easing the approach and the ability to grow and expand operational efficiencies. It also touches upon the fact of the 50% that's not been outsourced today and the growth that can come in that in extended lines of business as we look at adjacent plays and adjacent vertical plays that were touched upon earlier.

Lastly, there's a growing demand in integrated solutions. I'm going to give you a case example of that in a bit. An example of it here is rising fraud. Fraud has been prevalent in the space of payments in my space for years, $10 billion in fraud. And how we address that fraud and how we also address the solutioning for it is something that we've invested in. You'll hear about some of that in specialized services. You heard about it with Anish as well prior to that. And banks are coming and asking us, help us solve for the fraud tools. You'll also hear as an example of from specialized services, which is prevalent in my world in what we're doing in collections. And how we're leveraging AI and AI tools to make smarter decisions when it comes to collections and also those that take EI into account, all underlying with TP.ai FAB as the foundation.

As I keep peeling the onion, this is me. This is my vertical, $1.4 billion of revenue in 2024, 14% of the company's revenue. What you see is the microcosm of the key hubs for BFSI. We are 65,000 experts strong, and we don't take that lightly. That is what excites me, 65,000 leaders globally, the global approach with the local servicing, the local aspect that reminds ourselves of the regulatory environments by market, but also reminds us about the global investments that we can then bleed into and roll into the hubs that service our customer better.

With 350 clients globally, half of which are the top banks in the world, of which another 60% are the top 25 companies in the world, I'm humbled to service them day-to-day. And as Himadri alluded to, we're a one-office model. We not only think about a solution for front office and forget mid- and back office, we think about the customer and that customer journey many times resides along all of that.

On top of that, you layer on Akash's world and a Anish's world of data and AI and how we can service our customer kinder, smarter because we know more about them. And as you heard about on the panel as well and where our customer is in charge, they are in the steering wheel of making those decisions.

As I think about stepping back and saying, so how do you do that? You've talked about your strategy. You've seen double-digit growth over the years. You say you come from the industry, you're excited every day. So here's a snapshot of how my team and I, our 65,000 member team think about breaking down those verticals within BFSI. As you go even further, there's a modular approach. Anish spoke about LEGO pieces. I also love LEGOs. And I think the LEGO example is a very fair and a good accurate example of how we approach our clients. Sometimes, that approach is proactive. Sometimes that's through an RFP. But whether you're at the customer acquisition cycle of the customer or you're at customer retention or you're assessing risk, there's a gamut of capabilities that come with that, all holistic from a one-office approach that then create the differentiators that truly decipher us from our competition.

An integrated one office delivery model, which the team has talked about, sub-industry micro services that really create the nuances and the uniqueness of the vertical and the servicing that comes with that, strategic AI partnerships, some of which you've already heard about and you're going to see and touch and feel in a little while at break and the ecosystem that's built. Embedded analytics and agentic AI that is servicing and powering and is the DNA. It's my growth. It's what I have. It's the stickiness that I'm having with the customers, and it's the expansion and the wins that I'm having with new lines of business that are resulting in that cycle back to that modular approach.

Scaled agile and transformation, I said I was running it. So it's been around for a while, but the importance of that in the banking world and being able to pivot fast, have quick sprints, learn from your mistakes and pivot even more has been critical. And we are solutioning and supporting our clients on that front. And last but not least, last but not least, is the adherence to the regulatory environment and the audit infrastructure that's essential in my vertical.

So here's a couple of examples of execution. Thomas spoke about a large U.S. bank, one of the largest whose headquarters are right around the corner from here. I'm going to speak about another one. Another very large U.S. bank that we have been in -- had the privilege of having a relationship with for over 10 years. This was us from a core and building out the core platform with AI and us going proactively to the bank and saying, we have -- we've been working with you. We have investments that we want to copilot with you. We have a proposal of what we can do and our North Star are some of the impact metrics that you see.

We exceeded our North Star because they believed in us. It goes back to the trust and the relationships and the partnerships, and we exceeded the impact by showing 6 points in net promoter score, 15% reduction in handled and interaction time and an improvement of efficiencies to 20%.

That's exciting. And this was proactive engagement. That resulted in the growth that you see year-over-year of 12% in revenue, not only in the existing line of business that we had, but expansion into the expanded vertical plays because now we had trust. We woke up that morning thinking about them. As a customer, we also -- and that's an expectation of everybody on the team is you need to be a customer of the clients you're servicing.

In 500,000 employees in the company, there's definitely a customer in every single client that we have. And that's a secret sauce. To leverage the customer interface and the use case to say, I actually was in this interaction as a customer. And I have some ideas that I'd like to proactively share with you, not as a problem statement, but as a solution, and that's what we did here.

Another example of that is extending vertical plays. This was a start-up U.K. bank when we first started working with them. And they believed and partnered with us, and we believed in them as well to partner with them to start what resulted in them being the largest U.K. bank in Europe and growing. Their expansion plays has and already is expansion into the U.S. and a multitude of other countries. We've been their partner from the onset to grow with them and with them. And what you also see here is we've grown from care to an example of a conversation, a simple conversation with the Chief Operating Officer because as you grow as a bank, there's unfortunately, the unfortunate likes of fraud and fin crime that grow as well.

We partnered with them to solution on what we could do to service out of one of our centers of excellence, one of the key for fin crime and fraud in India, and we solution that as an expanded vertical play and showcase to them and you see their results. On both slides, there's also quotes from our clients, which I didn't cover, but it's at your leisure, and you can cover that when you please.

And with that, I'm going to hand it back to Mike.

So I was thinking about our client conversation and each of our clients had different problems, different opportunities, different things they were trying to solve for. Something that I heard is consistent throughout was TP as a partner, TP evolving, TP finding a way to create a solution for each problem over time.

We heard Miranda talk about the return to growth of BPS and how our solutions evolving with the market, integrating technology. Himadri, bringing the solutioning together, which seems wide, diverse, complex, but building a process flow that allows us to do it in an efficient, effective way means that we can bring that solution to each and every relationship. He also made the point that it evolves over time. It's not a one and done, which means we continue to solve and change different things with our clients and earn again those opportunities for new growth, new ways of working. And all of that is built on the 40 years of TP experience.

Now as we head out to break, I'll remind you again that we have some great demos in our demo rooms over here. And for those who are joining us remotely, we'll be sending a landing page so you can experience those demos as well and see how technology is changing the way work is done. So thank you, and I'll invite you to the demos.

[Break]

Okay. Welcome back. I hope you all enjoyed the break and even more so the demos, getting to see how we bring technology to help solve clients' needs. As we shift gears a little bit, we'll talk about our specialized services offering. And JC will walk us through each of those businesses, how they solve niche problems in high-value, high-important situations. Without further ado, J.C.

Thank you, Mike. First, let me present myself. My name is JC. I've been 20 years in the outsourcing business, 10 of those in IT, BPO and 10 of those in Teleperformance. And for the last 10 years, I've led the LatAm Teleperformance family. And then always seeing the specialized services as one of the most beautiful spots of Teleperformance. And then 10 days ago, I got a call from Thomas and Daniel, do you want to lead the specialized services for Teleperformance? Let me tell you, I didn't take 30 seconds to say absolutely yes. And I hope today, in the next 3, 4 slides, I'll show you why.

So first, we fit perfectly in the industry footprint that you were seeing this morning. We are an extension. We are very specialized companies that match TP's footprint and blueprint in every industry that you saw this morning. So let's talk a little bit more about that and about our AI transformation.

Let's start with AllianceOne, a collections company, fully embedded into BFSI, health care. We have been doing collections for years. We have had deep expertise that now we are leveraging with AI. So there's lots of synergies, lots of things that with the complement of the core of Teleperformance, we will continue to build an even greater collection company.

Then let's go to TLS, TLS Passport and Visa company. This is a full integration with AI and with the India and our TP.ai FAB. Imagine from the back to the front, all the automation that we are going through. In this post-COVID world in which we are open to travel, open to do business, open to get everywhere in the planet, the Visa is an amazing moment to live in.

Then PSG recruitment. Just think about how many people Teleperformance recruits. Think about having the strength of supporting the core and what of the clients think on the client footprint that we have in the core of Teleperformance. All those clients are going to hire us. This is a huge opportunity to continue developing in a very specialized niche of our business.

Then Health Advocate. I will not jump into Jeff -- where's Jeff? Into Jeff's thunder, but we have a beautiful employee advocacy company. Think about that. Think about having a specialized company in the health care vertical, the amount of opportunities and transformations we are leaving as we speak. This is going fast, and we are going to continue specializing ourselves.

Lastly, the interpretation companies. ZP, our latest acquisition. Sherri will later share a video, but our deaf and hard to hear and sign language business is beautiful. There's a lot of potential to leverage our current Teleperformance footprint with this. I'll share a little bit later about this.

And then LanguageLine, mission critical contact, mission-critical interpretation, imagine the amount of situations that globalization and contactability, the new use cases for this industry, the compliance, the regulation that will come to give a prosperous future to LanguageLine. So I'll speak a little bit later about this.

And then the addressable market, EUR 100 billion. Think of that. We have a huge pool to develop our business. We are through all this leverage that we do with the course of Teleperformance, plus the specialization of these businesses, we are going to continue to go even further in the growth and in the development of this specialized business.

So just some growth cases for you to think about. ZP expanding to new markets, the major markets of the world. The U.S. has 10 million and the global population that had -- that needs as to support them is EUR 70 million. So there's a huge possibility with the TP global network to expand this business by 2, by 3, by 4.

It's using the global footprint and capillarity of Teleperformance. And by the way, a lot to still be done in the domestic U.S. market. Then PSG, I don't -- Dave you'll later presenting our latest automation, but we built Ana. Ana is our AI platform that has already screened 50,000 people. Imagine that. We are doing part of Teleperformance recruitment and many clients using already Ana to screen candidates. This is developing fast. I would love if you take some time and you go and see the presentation and see it live. It will be great for you to build a better impression.

Then expanding to adjacent markets in our Health Advocate business, Imagine payers and providers. They are all Teleperformance clients. We have huge implementations. I remember in my LatAm days, these clients are huge. You see Thomas I am aged, but I have good memory.

We have beautiful health care clients around the world, and we are going to leverage and boost. It will shine and Health Advocate will be -- maybe the most beautiful star in the whole Teleperformance universe. Then, if this allows me, I will deep dive into LanguageLine. We have gone from voice to video interpretation to AI and think on the use cases, think of critical situations.

Go to an emergency in a hospital -- I'll share some examples later, and understand the person that's having an emergency, how he's feeling, how he's not using his native language and how he needs to be understood to really be able for the doctor to understand how -- what type of treatment does you need. Imagine a 911 call.

Now with globality, everything is interconnected. There's many new use cases. Compliance and regulation will take us to really do deep understanding in all the mission-critical contacts that the world is enabling. So LanguageLine will continue to grow because there's new use cases. There is new legislation and of course, AI interpretation will continue to boost the business.

So yes, 35 years of experience, our enormous footprint around the world and the knowledge we have on the business will leverage mission-critical contexts. I have no doubt LanguageLine will continue to be a booster of Teleperformance in the years to come.

So next, I'll share a video about what do we mean by mission-critical.

[Presentation]

Now we'll talk about ZP, our latest acquisition and how this expands our portfolio. For this we have Sherri Turpin, who have prepared a beautiful video.

Hi everyone, thank you, Juan Carlos. I'm Sherri Turpin, I'm the CEO of ZP Better Together. While I wish I could be with you in the room today, I am very honored to be able to walk you through the most recent acquisition of Teleperformance. We were acquired in February of '25 and we've had an amazing journey from when I started in 2015, and I'm -- we -- actually, everyone at ZP is very exciting for our future.

So I will tell you a little bit about ZP. Our mission is really clear. And it's one that has inspired me since I started in 2015. We are here to break down communication barriers. We are here to give the deaf community in the United States, which is where we currently are servicing, equal access that the hearing world takes for granted. It was astonishing to me how much the deaf community in the United States did not have per the ADA law just equal access and technology where they could live their life at work, at home and on the go.

So our mission is very clear. We are supporting American Sign Language, the community that's in their native language. We are really breaking down all barriers and providing everything they need to live their life to their fullest. We're very proud that 78% of our workforce is female.

A lot of people don't know we not only service the deaf, but we hire the deaf. 55% of our employees are deaf and hard of hearing. And we have over 4,000 American Sign Language interpreters, they're just amazing. A lot of people don't know that we're very proud 2020 through 2023, we've won the 100% score, which is the highest score for the disability equality index.

And we have 3 cultures here. We have a hearing culture, a deaf culture and an interpreter culture. And what really brought us together and made us be who we are today was really how we care for each culture individually and together, we are the powerhouse that has become the provider of choice in the United States.

A lot of people are unaware that Zee was the first provider in the United States to really launch the program that we service today, which is video relay service. A lot of people, I will be the first to raise my hand. I have never met the deaf when I started in 2015. I wasn't even aware that there was a program.

So I think before I get into where we want to go in our future, I think it would be really wise for me to just pause, show you a video that we put together. It took me a while, and I've had so many of my hearing colleagues and hearing friends go, "Oh, I get it. I now know what you're doing." So I'll pause. I'll let you watch the video and then I'll be right back.

[Presentation]

I'm back. So what do you think? A lot of people didn't realize it's a 3-way conversation. A lot of people didn't realize that the interpreter in the deaf or in video together. A lot of people didn't realize that the interpreter is literally interpreting in an audio call on a headset to the hearing party.

We are 24/7. A lot of people don't realize that we are their 911 if they choose us as their provider. So it is a wonderful business where we are doing wonderful work and to say that I am extremely proud that Teleperformance has acquired us and gives us what I'm now about to cover which is the future. And the opportunity for the future is so much brighter than the last 10 years that we've been building in the United States.

And that's because in -- internationally, there's over 70 million deaf and hard of hearing that need our service. And what's great is we have paired our future with where Teleperformance already is. And so their strong presence in the countries of Australia, Canada, Germany and France is just the immediate jump-off point that we are interested in, and we will be leveraging the relationships with Teleperformance and we will be partnering in the same way to bring communication access to the deaf in those countries through that government for video relay service.

We have built and know how to perfect in the United States. So all of those skill sets will be easily implemented in all of these countries. And I'm excited to get to work. I'm excited to show everyone, not only Teleperformance, but the markets and the deaf community that we are doing exactly what I told them we would do.

We would build in the United States. We would perfect in the United States and then we would find a partner that could help us grow globally, and we're doing just that.

So thank you for your time. If you have any questions, please do not hesitate to reach out. We are excited about our future. I hope you are as excited as we are. I hope you understand exactly how important and critical this service is for the deaf. Thank you for your time, everyone.

I will now turn it over to the CEO of Health Advocate, Jeff Cordell.

My name is Jeff Cordell and I am an unapologetic geek, and I'm excited to share with you a little bit of my background. I've been involved in some of the most disruptive communications events that have changed the way that we live and work today.

I hold over 30 patents in the area of cloud before cloud was cool. I also have patents in the area of digital communications and artificial intelligence. I'm excited to share more today. I joined the Teleperformance family with LanguageLine back in 2014 and then continue to help the specialized services divisions grow before I took on the role, as the CEO of Health Advocate, it was an acquisition back in 2021, and I joined the organization with the unbelievable responsibility and honor to be able to lead this organization.

It was 2,500 years ago that Confucius said, a healthy man wants 1,000 things and a sick man only wants one. Imagine how much has changed in the world of health care since that statement was made. The advancements are staggering and it's exciting what's happening.

But along with that, is the increase in health costs. Employers in the United States who are part of the health care and private insurance in the United States are seeing a rise between 8% and 9% in the 2025 year alone. And so I want to introduce a new topic. It's called population health. It's how do you approach addressing that increasing ramp.

And it involves a number of things, preventative and proactive engagement to take care of your own health and that of your family. And we're trying to bend that cost curve, and that is the basis of population health and it's where health advocate plays. A lot of new terms.

Those are staggering numbers, but let me bring it back to you who's sitting here in this room. Three things that matter most to you. One is your own money, you're paying for portions of your health care insurance in the United States, and we need to make sure that you get full value of that.

Number two, it's your health. A challenge probably every person has a health ring or a watch that they're tracking. And with access to data, it is easier, and it is smarter than ever. And the third one that's probably the most important of all is your family. There's no expense you would spare to be able to take care of your children, your wife, your partner, even your parents and in laws.

And this is the area where Health Advocate can provide services to that entire range of capabilities. Here's Health Advocate at a glance. Again, acquired and joined the Teleperformance team, the TP team in 2021. And you can see by the numbers, but let me give a really salient example of the type of experts we have.

Maybe you're new to town and you need to try to find a primary care physician. And you want to find somebody that's in your insurance plan, near your home and accepting new appointments. That's a challenge that our health care experts would love or maybe you had the worst news of your life, and you or a loved one was diagnosed with cancer. Where do I start? What do I do? How do I find an oncologist? How do I go through the process?

These are the things that we do. We have nurses, clinicians and doctors that are available to you are members. So today, our clients are employers and they buy the services just like they work with insurance organizations and employees become our members, and we have a large group of people that we provide a range of services, but it doesn't just end there.

I'm going to speak a little bit about mental health. If you want to talk about an unbelievably difficult and pervasive problem, not just since COVID, but boy, did it put an exclamation point on the problem that we need to address in mental health. We offer what has been called traditional EAP, employee assistance programs. You might be familiar with that. But more than anything, there's mental health solutions that we are investing heavily in this space.

And finally, there's a range of services that we provide that are for biometric screenings, that's early detection to try to identify somebody who has hypertension, who has a high A1c and those things with some preventative work, maybe a really inexpensive statin can avoid the emergency room or worse. These are the areas where Health Advocate perform services. But we're not done. Again, I'm the geek, and we are applying technology at a staggering pace in artificial intelligence.

And if you think that AI started in 2023, you're wrong, it's been decades. It's been used for a very long time, and I've got a couple of salient examples and case studies of how we're using it today. Our clients are also informing us about new products and services. Here's one. How many in this room are dealing with parents that are aging and they need help?

We have a product that we launched called Generations. It's been wildly successful. It's how do you the caretaker, the worker take care of mom or dad who might not be where you are. When do we consider alternatives to be able to help them through that period of time. It's a pervasive and growing problem today.

And then finally, I've got an example to talk about how not just mental health, but there's a correlation between mind and body and these are the new products and services that we are launching within Health Advocate.

First case study, a U.S. cyber tech company. I think 5,000 employees. When we add the rest of the lives involved, it's about 2:1. So about 11,000 actual lives when we consider the families that we support. And the point that I would love to take away from here is we created proactive digital engagement. This is before Agentic became cool.

The ability for us to have an outreach campaign on a new term that I want to introduce you to. It's called gaps in care. So we have ridiculous amounts of data about those members of ours, not just census information about the ages of your children and the programs that you're enrolled in but data -- claims data as well, and we're able to evaluate that.

So maybe you're about to turn 45 years old, and it's time for colonoscopy. If he doesn't do it, we can head off a whole series of these things. But we can find him and in the channel that is of his choice, we can text him and we will connect him to somebody who would be able to provide a colonoscopy. That is an example of closing of gaps in care, and we do it by the millions.

And so there's a whole series of other things that we did with this client with proactive digital engagement and the results were astounding. Population health, remember, that was the term that I introduced earlier. Literally, the person who wrote the textbook on Population Health, Dr. Ray Fabius had a third-party independent group called HealthNEXT and he followed this organization with us, and they have a 1,000 point scale to measure how healthy of a population this employer is.

The results were astounding. For Dr. Fabius, we drove 115-point improvement on a 1,000 point scale, he thought it wasn't possible in a single 12-month period using proactive digital engagement. Again, analytics to find those gaps in care and digital engagement to close those gaps in care. It was incredible. It was exciting and it's more to come. We have case studies that are available on this if you're interested.

I've got another one. This is a challenge, a large rail freight company. 50,000 employees, again, think 2.1, so about 110,000, 120,000 lives that we're supporting. And it's hard to reach this group of people, primarily male, tough guy. How do you get them to reach out for mental health? And we built a strategy with this company, right, in hand in glove with them. And the strategy was we're going to provide proactive engagement, those on-site, and we incented it and we identified a number of chronic disease conditions.

And at the same time, we provided information about available mental health through our services. The combination of those was astounding and you can see the immediate correlation between somebody's mental health. And in this case, hypertension was a chronic disease condition, very addressable. The combination of those programs was able to remove people from hypertensive down back into normal range. It's an exciting time, but I have this. It came across my desk just the other day, and I'm too excited not to share this real case study if you want to understand how artificial intelligence is improving our human engagement.

So here's how it works. Our EAP, our mental health, if you use our services, you can contact us either through web or chat or through our member engagement portal or the traditional phone. About 60% of the people elect to go see somebody face-to-face. If you're in city, all around New York, you'd have access. But there's 120 million people in the United States who are in mental health deserts. There is not access to those services. So we paired them with video, mental health services. Not only is it available same day next day, it's able to reach that population.

So here's where AI comes to play. I just want everybody to pay attention. It is a wonderful opportunity we've already recorded all of these calls for quality and monitoring purposes. That was the olden days. But now we use LLMs and we do speech to text, it's gotten so much better with AI and now we have a transcript of the entire 45-minute conversation and with really great AI engineers, you then create a summary that are the notes.

So why did we do that? Well, at the end of a 45-minute therapy conversation typically, you have to handwrite notes and transfer them to the case study. Well, now AI has made that so it's automatically done. It scribed to the notes and the results are astounding. 63% improvement in the reduction of post call note taking, mundane tasks. 83% of those texts are adopted, metrics that are fantastic. That means that it's basically done, and I might do some post editing, a little.

But the most amazing thing about the product and the service that we offered is the member responded. It felt like the therapist was connecting with me in a new and different way. They weren't busy taking notes or typing. They were more engaged. So it is easier. It is quicker, but the quality of the services that we provide are much improved. It's so exciting. I can't help myself.

So again, data is the lifeblood of everything that we talk about. Dr. Alyssa Scott is on my team as my Chief Analytics Officer. She has a team of data scientists and AI experts. Now we're not just making our -- in our system, great gaps in care closure we're now exposing it to our clients.

Here's an example of a dashboard of somebody who uses our EAP mental health services, and we can now describe the outcomes, the improvement if somebody engages in our programs that they got better. And if you're an employer, that means that there's a number of things that are downstream called presenteeism, absenteeism and ultimately, attrition.

Now we're ringing the bell of communities of why are we losing people when we can address services to keep that tenured intellectual property inside of the business. Dr. Ray Fabius talked about population health. It doesn't just improve the bending the cost curve. He has a direct correlation that the valuation of a company is tied to their employee population health. Lots to read about it. Dr. Ray Fabius is name, if you want to read some more.

So I'm going to close on this. Three things to remember. One is our clients today are employers. And we're super excited about the fact that numbers of Teleperformance, TP clients are able to use our services, and we're starting to see with Juan Carlos' help, JC's help that we're able to bring our services to those larger enterprise organizations. That's one.

The ROI is easy to calculate when you're able to keep people that are called high-cost claimants, but it's easier to calculate if you've been able to head off heart attacks and strokes, easy to calculate.

The second one is we are a high-tech human touch company, and I hope you can see by the couple of examples of what we're doing. Our nurses and doctors are ready and available to help you but we're using that same technology in every single one of our group. We call it Project Summer and summarization. At the end of every call, we have to type case notes, and we're shrinking it for every interaction, inbound and outbound. We're using that same AI power and magic that I described in that mental health scenario.

But here's the most exciting part of all of this. It turns out that the solutions we've created, which we call member engagement, primarily employees is completely transferable to other worlds, to the pharmaceutical world. So maybe you've launched a new drug, maybe you have results and you have to reach out to your patients, did that person that used one of the early detections for colon cancer and they're positive. Did they follow up? Did they share with their primary care physician? Did they get a colonoscopy? And do we have an oncologist available?

This is where we are our best. We're moving into those markets. And the one last statement that I would say is the payer insurance markets in order for them to have a HEDIS 5-star score, there's a number of patient and member engagement things they need to do, and we are excited about expanding our services to help them as well. So that is it. I believe I hand it off correctly to the man that needs a colonoscopy, maybe. I'm not sure.

A couple more years to get there. It's important to remember, as Jeff talks about all of the technology enablers within specialized services, those are all built in TP.ai FAB as one technology solution that underpins both of these large services, both specialized and core.

You also heard about a number of these services that have been integrated in our service delivery and TP core, whether it's the Health Advocate product that helps our employees take advantage of the benefits we provide, whether it's PSG with a recruiting solution that allows us to produce a better core product in our BPS division.

Additionally, by having these specialized businesses in a joint go-to-market, we're able to solve more problems for our clients, extend the relationship and deepen the business that we do.

From here, I'd like to welcome Anish back on stage to introduce our next session.

Okay. First of all, I'm going to invite Professor Martial Hebert from CMU to come on the stage and then Akash Pugalia, please.

So first of all, Professor Martial maybe, he doesn't need introduction for many of you. He is the dean of the school, which is ranked as world's #1 on AI research. So thank you, Professor for being with us. And thank you, Akash, for joining us.

My first question to professor, to you is that you would have seen in the last few interactions that we have done today and before that companies like TP are investing a lot on AI and Agentic solutions. And there is a lot that's happening in the innovation labs that you run. What we are deeply curious about is knowing what is actually cooking in those? And how does it impact companies like us in the future?

Sure. And first of all, let me thank you for the invitation and point out how happy we are to be part of the TP family and working with you. We were -- we started a human computer interaction department, a machine learning department, language technology department before any of those things were popular. And from all the presentations this month, those are all very relevant to what you do. So I'm looking forward to this partnership.

To answer your question, let me point out a couple of things and relate it a little bit to the history of AI. So along the history of AI, we came across a number of solutions, a number of approaches that at the time were labeled as being the solution, okay? And this is it. We solved it, right? We just need to perfect it a little bit. And then, of course, what happens is that the next approach, the next revolution, the next solution comes along.

This is going to happen again. We have heard a lot the world LLMs, transformers. This is assumed to be the solution. This is going to change and very quickly. Let me give you one example. One of our faculty members in machine learning Albert Gu developed an architecture called Mamba, which is a possible replacement to transformer with much higher performance and more importantly, much higher efficiency, both in terms of computing and in terms of data. So this is just one example of what I mean here, and that transformation will take place very quickly.

The second aspect that I'll point out, and again, going back in the history of AI, this time, I'm going to go all the way to the beginning, 1955 -- was the first AI program, the Logic Theorist, which was done at CMU by our founders, Allen Newell and Herbert Simon. Why do I point this out? This program, the reason why it was significant, it was the first time that somebody looked at computing, not just to crunch numbers or organize data, but actually to do reasoning and work with symbols.

And in fact, for many decades, AI was entirely about reasoning and symbols. And then we moved to data-driven techniques. And we moved all the way to the other extreme, Well, now there's no reasoning. There's no symbols. There's only data in statistics.

The exciting thing is that now we're moving back a little bit, actually a lot. And we're now reintegrating those concepts of reasoning and knowledge and symbols into AI. Why is that exciting? It's because now there is an opportunity to not just look at a black box and look at its output. But actually, things like people have call the term chain of reasoning, for example, if you look at the most recent OpenAI development, for example.

That gives the opportunity to have much deeper output from the AI system. Something that may not be completely evident, that also affects how we think about AI safety. One aspect of AI safety has to do with explainability, which is totally inexistent, let's face it, on black box system. But now having that kind of reasoning output allows us to go to that explainability.

Now one issue with those new approaches bring in knowledge, et cetera, is that they may be limited in terms of computation because they now require additional computation. In your world, for example, where you might have issues on guaranteed delivery time. For example, that becomes a problem. So there is another thread and another wave of development having to do with any time computation, et cetera.

And finally that development allows a lot more integration of domain knowledge. There is a vast amount of domain knowledge in any domain, any of the verticals that you indicated. That is not necessary to learn from data. That can be integrated directly except we did not have the formal method to link those two things: One, symbol reasoning, et cetera, on the other side, data statistics, et cetera. So that's the exciting development that I see in the future.

Thank you. And with that, I'm very proud to announce we are joining hands with you to build the next level of research and bring some of the advanced research that you do for the benefit of our clients. So thank you for joining us in that partnership. Akash, I'll go to you what you heard from the professor on how the new advancements are happening. What specific role do you believe that data will play and in turn, TP can enable that to take the advancement on the research forward?

Thank you, Anish. It will be similar to what I mentioned earlier today. Data is the foundational layer in AI. If you think about models, how much ever sophisticated the model is, it's -- the outcome will not be relevant if the data is not accurate. If the data quality is not good. What we need to look at is how the fine-tuning of the model is happening, what sort of data quality that's there. That's where TP come into play. Now TP is at the juncture where we are implementing how humans are training the AI over a period of time as well as AI is training humans at the moment, right?

That is something that we are embedding in our operations, that cyclical process, it's the feedback loop mechanism that is helping our clients over how we can provide inputs to the large language models. That is where the orchestration layer is happening where humans' expertise as well as machine intelligence are coming together. And that's where TP is already playing. TP is also embedded with a lot of our clients in their AI development life cycle, so not only model training, model fine-tuning, but also looking at prompt quality and so on. So we are embedded throughout the life cycle and able to answer multiple questions that they are looking at through our highly specialized skilled resources. Yes, I mean, it's throughout the value chain.

And Professor to build it further, Akash spoke about teaming up of AI and the human expert. I understand you've done a lot of research on that personally and your university has, too. So how do you see -- what makes them effective as they work together? Is there more new frontiers to be established?

Yes. So let me give a little bit of background regarding CMU in particular. The school of Computer Science, which was the first of its kind, has 7 departments. And interestingly, one of the department is human computer interaction, which is where most of our interaction is, by the way, and here is the interesting thing about this department. Even though it's under the school of computer science, most of the professors in that department are not what you would think of as core computer scientists or AI scientists. They are designers. They are social scientists. They are learning scientists, not in the sense of machine learning scientists but learning science education. And coming together creates that new field, if you will, of computing interaction.

Now to answer your question as to where that's going. One of the transformation over the past few years that I think we have not completely appreciated yet. It's the degree to which the interaction between the AI and the human, it's completely transformed, okay? It used to be -- and we did a lot of work for decades on AI, human teaming. But honestly, what we call human AI teaming was you have this AI program here. You have the person here, the exchange information or the one give input to the other and gets the output back and so forth.

The level of interaction that we're able to do now is unprecedented. In terms of the behavior modification on both sides, just like people interacting really, except that now I'm not interacting with you as a person having a person's personality but I'm interacting with an AI agent, having an agent personality. I try to avoid, by the way, using anthropomorphic adjective to AI. But I'll make an exception for that one. So there is an opportunity to do the following: To go much further than just having AI human interaction in systems and to embed in the engineering, but to create a whole new discipline. And to answer your question, that's what we're trying to do, to create a whole new discipline at the intersection of AI, psychology, social sciences, et cetera., do formal models of those interactions and be able to use those now to make progress in designing those systems.

Thank you. And allow me to push you a little further on that professor. Companies like us are the practitioners of the dynamic interaction that's happening between the AI and the human experts. In your belief, what is it that companies like us need to do right or avoid the risk of it going in the wrong direction?

Well, part of the direction is what I mentioned, which is the formal modeling of the human interaction and couple that with the AI models. In other words, not seeing those as 2 separate -- 2 separate pieces that come together, but see that as an integrated piece from the beginning.

Yes. So Akash, as you see, these interactions are resulting into a lot of data. Agents are generating data that's being used to train them even further, human experts are generating data that's being trained, that's being used to train the AI even further. What role do you think that we can play or companies like us can play to make the best use of data to train AI in the future?

Yes. I think there's one fact that I don't think a lot of you would know. So I was reading it yesterday. In the next 3 years, the amount of data that will get generated will be more than all the data that has gotten generated in human history till now? So you can imagine the amount of data that will come in. And that means that the AI cannot just manage all the data on its own, the data needs to be -- again, the data needs to be fine-tuned, it needs to be high quality, it needs to be curated. And that is where TP comes into place.

So if think about agentic AI, we're already embedding ourselves in the vertical solutions, in the domain-specific solutions that are required to look at that data to define what the capabilities and outcomes should be. So that is where we are playing. I think another thing that the professor touched upon that we are already doing is the cultural integration of AI with humans. So that's something that, while as consumers we're using it. But from a B2B standpoint, how we are ensuring that our agents are ready to use AI in helping our clients, that's the other big wave that we are driving.

So yes, I think there is enough and more that we are doing for our clients in this. The last point that I'll say is that, again, data is at the infrastructure layer. We, at TP, are already playing in that infrastructure layer at scale.

Thank you. So let me just turn the needle in a different direction. There are a lot of investors in this conversation. And they are obviously evaluating companies like TP on the investments that we are making, both on AI and transformation. For their benefit, would you like to elaborate on what are the indicators or the trends they should observe to really ascertain that the investments are going in the right direction, and it's a sustainable competitive advantage?

Yes. So I'll concentrate on one aspect, which is what I call the AI measurement. And in fact, this is related to something that was mentioned earlier. I think Thomas mentioned this, the checking AI basically. And this is one of the hardest things and one of the key things to differentiate different approaches and different implementation of AI, meaning to what extent we are able to make measurement on that system that are informative of performance. If you look at a classical engineered system, any of the systems that we use every day. The reason why we can trust them and the reason why a company can make progress in developing them is that there's approximately 200 years of engineering science behind it, meaning things that range from formal mathematical method to statistical method all the way to experiment or break based practices, right?

We don't have any of that or we don't have most of that in AI. Even something as simple as being able to define the right metrics and the right measurement protocol to evaluate AI system is not something that we have at least formally. So that's something that we need to look at, and that's something that is a key evaluator, a key evaluation feature for any AI development and AI system. Now just to make it clear how this can go wrong, even in very things that can be -- feel very obvious. Let me give you an example for my work from computer vision, right?

One of the things that one needs to do in computer vision system, for example, for self-driving car and things like this, is to detect objects and to label regions in images, right? So you want to say that this pixel on the person, this pixel is on the car. An obvious measure of how well that works would be to say what percentage of pixels do I label correctly. That sounds obvious. But in fact, if you think about it, what application, what scenario actually requires 100% of the pixels in an image to be labeled correctly? The answer is none.

You never need to have every single pixel labeled currently. So this is an example where by using this measure, we're using the wrong measure for the task, clearly, but we're also using one that is much harder than what's actually needed for the test. So you see even in this case, when we think, well, it's obvious, that's how I'm going to evaluate my computer vision component and maybe how I'm going to decide to make investments because the one that is better in that metric is probably better to pursue, right? That's actually the wrong measure. And that's true across the board actually.

So this is a key, I think, a central point to concentrate on is given the -- what we're trying to achieve, given the context, the application, the verticals [indiscernible] what are the right measures, internal to the AI system. And one last thing on this in terms of you were asking earlier about current developments, right, it's maybe very difficult to define those measures explicitly the way I describe it, number of pixels [indiscernible]. So there's a field that emerges now with its own conference, it's called experiment or machine learning, right, which for people like me, older people who were raised old style in science and technology is an atrocious thing. Why would you have something that you created now need experiments, right? But in fact, it makes sense in the sense of using protocols to be able to measure those machine learning systems and to have measures that can actually be used to evaluate them.

All right. And Akash, anything further that you would want to add on that?

Nothing. I think the one point that I'll just add on to what Professor said is the feature on responsible AI and ethical AI is also an area where there is no consistency. So responsible AI is about bias mitigation, about accurate data, about hallucination and ethical AI is about governance and ensuring regulatory compliance and that is something that is not nonstandardized. And I think as the field is developing and growing, I think it will become more -- it will come more and more focused on what we can do as service providers to help clients in that situation.

Thank you, thank you, both. I'm just going to add one comment. Our intent of doing this partnership with world's #1 institute is simply very, very easily understood. We are going double diving into our overall TP.ai FAB. And this partnership allows us to do a lot of research on possible industrialized solutions that we can take to market much faster than any of the competition. And thank you, Professor. Thank you, Akash, for joining us today.

Thank you so much.

Thank you, guys.

Super insightful discussion. And I think it's interesting, as I reflect we can all probably imagine how computer to human interactions have changed even in our time, where most of our interactions were done at a keyboard to being done at a smartphone or some kind of tablet to interacting via voice with technology systems. And we can imagine in the future that become -- that technology becomes more and more integrated in how we live our daily lives. So being at the forefront of those new technologies, understanding how we build them into our service offerings, how people change around that technology will be super critical as we go forward. I'd like to invite Olivier, our Group CFO, to the stage to take us through the future outlook.

Thank you, Mike. Good morning to all. I'm going to be much more down to earth to explain what are our outlook. You have heard our colleagues all today, all this morning and the plan that was presented to you for the next 3 year. We translate this plan in figures that I'm going to share with you. So I wanted to reiterate the financial ambition that the group has set for the strategic plan, sorry, over the next 3 years, ending at the end of 2028. First of all, growth. I would say 3 things of it. First of all, we are setting a target for 3 years. For those who remember the years before, they were asking us to deliver our guidance for more than 1 year.

So we are delivering a 3-year guidance 4% to 6% third point. Second point, we are setting this target at 4% to 6% mid-single digit by 2028. And I just wanted to hammer that. This is a significant performance. It means that either in 2026, either in 2027, you will see an increase of like-for-like versus of 2025 guidance. And I just wanted to be clear here, we are going to continue to grow in 2026 and in 2027 before achieving hopefully the high bar of this guidance for 2028. Second point, operating margin. The objective here is to continue, and I'm using this word by -- not by chance, to continue improving the margin to reach an operating margin on revenue of around 15.5% by 2028.

Of course, and you understood that from all the days that we have today, we will make significant effort over these 3 years to build, develop and enrich its AI platform that was presented to you. These financial organic effort, either OpEx and CapEx will be largely offset by the area of cost reduction initiative that can be grouped around 3 priority action. First one, the completion of the implementation path of the synergy plan following the merger with Majorel. The implementation of the plant in France that has just been finalized or started finally to start will enhance the competitiveness of our operation with effect from full year by 2026. Second one, furthermore, in advance, and you understood the GC is there to help us. And advances integration, an internal synergy plan for specialized service activities with [indiscernible] will be implemented starting as early as second half of 2025.

It aims to pull several costs with shared service, but also systematically rely on internal resource for all offshore interpretation activity as an example or further integration for LLS. Finally, you heard a lot about generative AI to sell, to develop our business. But generative AI will simply -- will greatly simplify the group organization by making it more fluid and more integrated across its various functions, thus lowering the amplification and delaying of organization. So clearly, you understand there is a clear set of plan of savings that are going to largely offset the investment, either OpEx and CapEx that are designed to develop the business today. Let's move to the balance sheet structure. The group aims for net debt total to recurring EBITDA ratio of around 1.2.

TP has the strongest balance sheet in the sector. We wanted to rub it further. The group intends to strengthen the situation by deleting a significant portion of our cash flow to this debt reduction. Finally, and I wanted to be clear on that, the group is going to deliver EUR 3 billion cash flow over this 3 period. Just looking at this figure versus the market cap. In free cash flow versus period '26-'28 offering for substantial potential return to shareholders while maintaining a strong financial discipline and investing in the future. This is the case.

So this group is going to generate EUR 3 billion to deleverage and to return cash to shareholders. How we are going to make it? What is the planned capital allocation over the next 3 years? 20% of this free cash flow roughly around EUR 60 million will be allocated to additional investment to accelerate the group AI transformation. You understood that we are working with a lot of people, either partnership, small bolt-on acquisition, speak with a lot of people to enhance this AI transformation across the group to be the leader. This money will be spent over 3 years.

This is a basket that we need to adjust precisely. But this is something that we need to do on our capital allocation to continue to drive growth. As you understood, 30% will be used to further strengthen our balance sheet. We will finish with -- at the end of this plan with a debt-to-EBITDA ratio of 1.2, largely better than all the competition. This is something that matters for the Fortune 500 companies to work with.

And lastly, 50% will be returned to cash to shareholders. We are speaking of EUR 1.5 billion. It's not small money. There are still discussion about how we are going to make it between dividend policy and share buyback. But at least we will maintain the dividend policy. There will be share buyback. And this EUR 1.5 billion will be returned to shareholders over this next 3 years. Of course, the group wish to retain certain flexibility for any strategical and financial value creative acquisition. But what is clear today. It's not our group intention to carry such an operation in the short term. That's just what I wanted to let you know. And now I hand over to Thomas and he is going to conclude -- to make the wrap up.

Thank you, Olivier. So maybe to sum it up and before we go into the highlights to really try to give you the big picture from our perspective. If you think back what we've heard the last 3 hours, we're trying to paint a picture for the future. And what is our view as TP about the future, not in 15 years, who knows maybe by then we live in the singularity if Martial will develop the AI. But what is the future of the next 3 to 5 years? Where do we see the business developing? Is our view that AI will replace humans and basically take the human interaction completely out of the picture and act autonomously or is our belief, yes, AI will be ubiquitous but it will be, as I said before, like electrical current.

It will supercharge the work we do, enabled by humans working together in a new form together with AI. We are a B2B services company that worked in complex enterprise environments that have to fulfill security data, responsible use of being specific to the use cases we see in our clients and drive outcomes for them. Our core belief from everybody on stage today and when you see also recent articles in the news is not that AI will take humans out of the picture, but it will change the way we operate. It will make us more efficient. It will make us more effective. It will create different relationships between human and powered by AI and agenetic AI, but it's about the future we see for our business that enables the interaction with both.

The history of TP over the last 40 years was a history of change, and this is what we're seeing in the next 3 years. Changing our core with our strategy, expanding our value chain and capturing the new opportunities you heard from Anish and Akash, but we do this always with the same for ingredients. Managing people at scale, which is a key component in today's world. It's not an easy task to do this around 100 countries worldwide, being process excellent, being obsessed with process consistency, adherence and excellence performance. Having the technology know-how, you heard about the 3,000 people we have, driving this change partnering with the best companies we see in our [indiscernible] today. These are the world-class partners. We are happy and we are privileged to have in our site and having domain expertise. You heard from our colleagues, Mamta spend many years in the banking industry now joins us with her 65 colleagues to drive that domain expertise. And as a company that is capital light, we don't have large assets, we don't have large factories.

We are agile and can adapt to these opportunities quite rapidly using our capabilities that we have and drive this, and you saw this in TP.ai, the foundation of what we're trying to do, intelligent orchestration based on secure infrastructure of human and AI for outcomes, for industry-specific outcomes that matter. And there, we do believe we are the best partner for our clients, for our business partners that we have, for our employees and at the end also for our shareholders because this allows us in an efficient manner to deliver significant cash flow, yes, that we will responsibly invest in our business to grow it, and we see growth opportunities in all 3 strategic avenues, but also to return it to you because that's our commitment as management to put the money that is not necessary to develop and transform the business back to our shareholders.

With that, I will hand over for the closing remarks for our legendary founder.Daniel, over to you.

Thank you so much. I hope that you have seen that there is a lot of positive energy in this group and a lot of talent. That's the first point. The second point I'm going just to say 2 things about AI. First, data, garbage in, garbage out. And there is something very important today is that there is a bottleneck in new human data because everything has already been integrated by the LLM and clearly, there is a need for reinvestment by the humans of the data that can be generated by the AI.

Second, yes, we leave more and more in the digital world. And our kids even more than us. But you know what, the digital world is as dangerous as real world. And it's even more dangerous because there are less cups. And yes, we do a job that I'm super proud to do which is business integrity and trust and safety. Business integrity is against a deepfake and trust and safety is maybe to protect your kids from the pawn or whatever. By the way, I would like to invite you to read what the new American Pope said 2 days ago. So not only we use AI, but we work on AI. We enrich AI, we control AI, we transform AI. So it's a full partnership. We live in, with, besides, inside, we are totally integrated.

Now what I want to say also is that the beauty of the world is its diversity. First, companies have very different strategies depending on the product, the service, their positioning. Second, the reason to interact for an individual with the company are multiple. Some are mundane. Some are extremely important emotionally speaking. And you can have a lot of dissatisfaction that can be created. I remember when we started to do customer service in the '80s, most of you were not born.

There was something that we knew already which is a satisfied client is going to impact 3 persons when dissatisfied client is going to impact 10 persons. So our key role is not to be just a facilitator but a loyalty builder. Finally, yes, all the population are not the same. They have different ages. And of course, if you are born after 2000 maybe you spend your time on -- I'm not going to mention the name of our clients. But you also have an aging population. And by the way, Jeff was saying, yes, you even have to take care of your parents. Can you imagine? But you know this aging population prefers to have a human contact. And think about what it means statistically.

So it's not one size fits all. It's super important to have that in mind. And this goes also with the education level. I can tell you, we can make a lot of correlation, a lot of segmentation, correlation and find different ways to communicate to different public for different kind of topics. And that's why because we live in, within, besides AI and at the same time because the world is diverse and because the world changes every day that we have a great future. Thank you.

Now we'll open up to questions in the room.

This is Suhasini from Goldman Sachs. I have 3 please. One is on the top line growth, 4% to 6% like-for-like growth in 2028. I appreciate the color you gave on your expectations for '26 and '27. But given AI potentially has deflationary effects on the pricing, how do you see the pricing versus volume component in your target when you set out that for 2028? And do you have any expectations on growth by division, specialized services versus core services and DIBS.

Maybe I'm going to try to answer and then as he is much smarter than me, he is going to finish the answer. Really you are right, AI has deflationist impact but AI generate also new needs. So it is super complex. You can imagine it is super complex to make a model. So what we build the future depending on what we see already today. And what we see today is that the kind of much more engineered like Himadri was presenting. But at the end of the day, the revenue that we generate with our largest clients tend not to be deflationist. There is more volume. There is more geo strategy. There is more digital integration.

But at the end of the day, we see a kind of stability. But what is important is that, yes, it's important that we open the new line of business like Anish -- no, Akash was presenting in data annotation like a very traditional business that nobody mentioned and that is but massive for Teleperformance, which are the B2B sales because we are passing from an ecosystem in which the company used to own their technology to an ecosystem where the company go to cloud, but cloud generate much more interaction than when you own a technology. So the numbers that we give are a mix between our business development plans, what we see with our largest client and the fact that, yes, we consider that we are not going to be able to increase the price per interaction because we are in this very, very transformative environment. Thomas?

So, to add to what Daniel said, it's, I think, too simplistic to just look at volume and price per interaction. You have to look at the line of businesses we do provide. And that's when we try to build a model for the future and to resource for the respective areas, where do we see growth we want to invest. As you see in our numbers, we have roughly 6% or so of our revenue is sales, for instance. And it's a line which is not a cost factor with companies but a revenue-generating business, and we see a lot of growth momentum in B2B sales. Same thing is true for our data services, which is reported within trust and safety. It's a line where we see a lot of potential. And there, we will see -- it's not a question of interaction, but opportunities to provide that service.

Another example is our back-office services line, where we extend and what Miranda and Himadri presented, to see opportunity, how we can grow further high-single digit in this area of back-office services. So it's really line by type of business, same thing health advocate, you saw the presentation, recruiting services with P&G LanguageLine, ZP, where we see where do we see growth and where we want to double down on that growth. And yes, technology will make in a status quo the price per interaction or the cost per interaction less, but that's been the case for many years, and it's about adding additional layout services solution on top of that.

And we have tried to be careful.

And any color on growth by division, core services versus specialized services?

I can repeat what we said in our Q1 numbers. We're very happy with the growth momentum we see in our core BPS business. We are a little bit cautious for the growth perspective of our specialized services as we have this special environment right now for LanguageLine. But in both areas, we're super excited and as Juan Carlos said, we have beautiful clients in both divisions, and we want to serve both of them.

I would say one thing. We see our core service business growing stronger than what one could have expected or what could be the market feels expressed in '23 or '24. Now it's true that we have a specific, difficult timing directly related to what's happening in this country and that everybody can understand. That's for the specialized service and specifically for our largest company. Now our largest company is in great shape. The margins are not impacted. And in fact, we think that the kind of stabilization or lower growth that we see right now is something that is going to be temporary. We, in fact, had this phenomenal already, not at the same scale in early 2017, and I'm not going to comment now.

Just moving to margins, 15.5% margins. It's 50 bps expansion between 2024 and '28. Given AI can potentially give you margin accretive revenues, I think it was mentioned in one of the presentations, why not look for more?

Of course, we could have bragged for more. What we have designed in our model is, of course, some costs associated to that, some savings. I tried to explain what kind of savings we will do. So we have been putting that in a model to see how it's going to be phased because this is a problem is a phasing. So it might change small from 2026 to 2027. And finally, we thought that lending at 15.5% was a good promise. We hope to be better, but we thought it was not really needed at this time to bet on the future. We are reasonably confident that it could be higher, but let's do it. Let's deliver it first.

Yes. There's no free lunch. So of course, we have to invest to reap the benefits.

Got it. And my last question is on the free cash flow, please. EUR 3 billion of cumulative free cash flow by '28. You generated EUR1 billion in 2024. You're planning to grow your top line margins over the next 3 years. Why is your free cash flow not more than EUR 3 billion. What am I missing here?

Conservatism again.

No, it includes also the internal. So if we invest, for instance, in the internal teams, you saw the strategy from initial cash. We're investing it. We're investing, as we speak, in our B2B sales team. So we do internal investments, and this cash flow is net of those.

Is there any change to CapEx to sales guidance?

No, no, no. dramatically, the CapEx will be between 2 and 2.5 probably close to 2.2. It's not dramatically changed. It might be early to jump by 10 basis points or 20 basis points a year on other, but it's not going to be dramatically changed.

My name is Soren Shaw. Thank you for today and the demonstrations Two-part question. You made the comment earlier about 1 size does not fit all. And we saw a demonstration in the back with 1 of your travel digital agents. And the demo gave the example of the cancellation of reservation, switching dates. What do you think in terms of segmenting the AI offering. So is there -- do you think there will be one base level of functionality? And then if you want additional functionality for that AI tool, there will be a pricing mechanism.

So you'll have opportunities to collect lots of different revenue off of one offering based on functionality. A little bit like Bloomberg does, right? You can get one layer of functionality. And then if you want additional data pieces you pay extra? And then second question is with the human agents are very good at cross-selling, how do you envision kind of a digital agent doing a cross-sell?

Thank you for the question. The first question is a very easy answer, absolutely, yes. There will be multilayering of the time of agentic AI the client wants to consume. So as the consumption increases, the price increases and the good part is, and it's Olivier's quote but I'll say it anyway, the more they consume, it is GM accretive. So therefore, it adds far more value to us if they continue to buy more. And the likelihood of them buying more from us is even higher because the customer outcomes are far more greatly impacted. You saw that in one of the cases. And before we could demonstrate, there was supposed to be an outbound call, which would have registered the customer for alternative restaurants.

So this is an additional service they could buy. So absolutely, yes. We believe that a lot of this design that's going to happen is going to get ported across many other industries. So whatever you could do, for example, an insurance claim that we earlier spoke about can potentially be transported to health care claims. So there will be that portability that will happen, and you could take one case to another and then make margins out of it.

There's another question.

Core services has already started to rebound, but it's going to take you guys another 3 years or so to get to the mid-single-digit growth rate. What's going to happen in between now and then that's causing that?

So maybe there's a misunderstanding. We're seeing progressively improving. So as you know, the guidance for this year is 2% to 4%. If we adjust of this nonrenewal, as you know, from Specialized Services, would be 3% to 5% and we see the growth rate progressively moving to this mid-single digit.

Yes. I don't think so that we are going to need 3 years.

I am Nicole Manion from UBS. Just one question from me, actually. Olivier, I think you already kind of hinted what you said on the CapEx to a percentage of sales outlook. But could you maybe talk a little bit more about what you see as kind of the split within those AI investments between sort of what runs through the P&L, what might be additional CapEx? Any additional kind of detail on that?

Difficult to give you a precise figure. But clearly, we do expect to have some additional CapEx in -- not in a big size, but I'm speaking of the CapEx, not on the investments that are -- could be used for the free cash flow. I just make the difference. But I don't expect something dramatically higher than EUR 100 million, EUR 150 million max.

There is something very important already. You know that we pass in our infrastructure, we pass mostly to cloud also. So we significantly have less CapEx than before, but it means that it has been translated somehow into OpEx.

You move from depreciation to subscription cloud.

Other questions maybe from the remote business?

Fantastic. Markus Schmitt from ODDO asked you said you aim for a net leverage of 1.2 by financial year '28 from 1.9 most recently. So what rating do you aim for then?

Today, the group is rated BBB, as you know, Standard & Poor with no precise target, I would say, let's put it this way. We have no public target. We might say that if you have 1.2x EBITDA -- recurring EBITDA debt ratio, you would be at least BBB+. That's what I can say.

An analyst from Kepler asked, the group's pivot toward data services and consulting will expose you to a new competitive landscape. Which companies do you expect to increasingly compete with TP going forward? Could you provide some examples and the key competitive advantage compared to those players?

I can try to start to answer, then as usual, we are going to make our view. First, the borders become blurry. They are blurry between BPO and CX. They are blurry between IT service and BPO and CX. The low code, no code changed many things. And it means that the competitive environment is moving. Yes, they are the direct competition, but I would say that the direct competition maybe there is one other player that present characteristic, more or less similar to TP. Otherwise, we should take advantage of this situation because we are in an ecosystem in an environment of consolidation. And really consolidation means that we have more opportunity to grow and to be successful. Now I prefer not to mention any companies, but yes, we can be in competition for some line of services with large companies that present themselves like consulting companies, but that are also BPO companies. .

We can be also in competition with IT service companies, with BPO companies. I would say that right now, I think that there is 1 geography. And strangely, I'm not going to speak about a nature of companies, but I would say that there is one geography that today present a lot of positive characteristic, which is India. And fortunately, our largest company, where we are the stronger is in India. And clearly, I would say, our most direct competition are going to be several Indian companies.

I would add, if you think about the markets we described, yes, we have the trend in the core BPS business of consolidation. And yes, with our new offering, we are entering new markets, but we do have 2 distinct advantages. One is these markets, in particular, when you think about data services are just evolving. So it's an open space that exists. You have newcomers coming in startups. You have large established players trying to make the venue, but TP comes from a position of strength, giving our global network, giving our existing client relationship and giving our process capabilities.

So it's more a question about execution that we see winning when you look at the pipeline and winning that the deals we have at hand, improving our excellence operational delivery in these elements because we do come, I think Anish talked about it before, from the actual floor from the operation where the AI is at place. So we don't have a DNA coming from a PowerPoint and then trying to make it happen. We come from a status that we are already in the midst of making, bringing the AI to life talking to the interaction, managing the processes and then basically moving up the value chain.

So from that perspective, I do believe we have a lot of strength and tailwind with us. as we're actually coming from the actual operational floor.

Maybe one other point on that. I would say that our competitive advantage is the diversity and the stickiness of our client base. It's our kind of distribution power. And that's the reason why many new gen AI companies are very interested to partner with us because somehow we have a fantastic advantage, which are the decades in which we have been serving our clients. And we have a deep trust relationship with our clients.

Mr. Schumacher asked, why don't you stop the dividends and focus on buying back your shares. The return on your shares is higher than any other options out there and higher than the AI investments you're making.

So first of all, the question is not raised for 2025. You understood that 2025, we made the acquisition of ZP. So the question will be raised again in 2026. What we have set up as a target was to give back to the shareholder EUR 1.5 billion over 3 years. First point, I'm sure for those who are not perfectly aware, but buying back share cost on top of the price that we pay 8% of tax. I just wanted to remember that this is something that doesn't help. Decision of how it will be exactly split between dividend and share buyback has not been made. Board will be probably decide that later on. I've had time to see that such a company like us will catch the dividend announced to the market that will cut the dividend. I don't see that happening, clearly. So to have a dividend -- a reasonable dividend over a period makes sense, makes sense for most of our shareholders, not all of them, but I know there are some that are just interested in share buyback. We do believe that dividend is also part of the story. Again...

And we can say also that the opinion on this topic are very different.

Yes, both sides of the Atlantic. Again, I just wanted to -- just to highlight the point. I have not made the math, but I believe that the free cash flow yield should be beyond 20% as we speak. I maybe asking you -- both of you on, but you are much more closer than 25%. So I know companies are in difficulty, but companies that are 25% free cash flow yield, I've seen worse situation in my life.

Great. Rami Grenu from Morgan Stanley has 2 questions. Margin guidance of 15.5% after AI transformation to confirm. Can you help us understand what level of margin you expect through this transition? Could it decline from 15% you're targeting this year?

Look, we provide today the midterm guidance that we provide the update for '26 and '27 in the respective guidance for the year, as Olivier indicated below, we are investing in the business, but at the same time, we are also investing in efficiencies that should offset the investments, and we will provide more details as we move forward. But we are committed that every efficiency we gain is, of course, translating higher margins, including the investments we do.

And the company is totally committed in these 3 plans that I just mentioned earlier on that are starting. Some of them are already launched. Some other are already started. So we are absolutely convinced that we will be able to deleverage our company and to improve our SG&A and direct costs on top of that.

And second question, what led to the change of management within specialized services?

So it's very simple. Scott Klein has been a fantastic leader for specialized services and specifically driving the incredible growth of LanguageLine. I would like to remind to everybody that when we acquired LanguageLine in 2014. It was less than half of what it is today in revenue and of course, profitability. It was very interesting because at the time, there were many people who were telling us that we were crazy to acquire LanguageLine because that would be a business that would disappear in the following 2 years. That was 11 years ago and LanguageLine is in very good shape. So Scott Klein over the last several years already was considering to retire and he made the decision to retire by the end of July. And by the way, the recommendation of Scott to us was also to have JC. In fact, his real name is Juan Carlos, to be the 1 who will take the succession.

Thank you. [indiscernible] from Berenberg asked guidance of MSD organic to 2028 implies that '26, '27 will be below that. Could you please clarify what you mean here?

Say it again, I'm not sure -- when you say that the 2026, '27 will be lower than '25? That's what he said?

I believe, lower than '28, if you could clarify.

No. Again, I'm not going to answer on detailed guidance for '26, '27, the budget process has not even started as we speak. We have, of course, plus and minus, and it has been mentioned both side by Daniel and by Thomas. We are working on it. Of course, we are a reasonable company. We will adjust to deliver something that is -- that will be in line with the market weighting.

I'm going to try to give a conceptual answer. We are in the range of the marginal variation and it's extraordinarily difficult, 3 years in advance to come and to say exactly what will be the plus and the minus in '26 and the plus and the minus in '27. So we decided to give a commitment for '28. We do not expect to have something material impacting our margin in '26 and '27. And I hope that this conceptual answer clarify the things.

And last question from our online attendees, Rob [indiscernible] Investment. TLS has contributed nicely in the past, but seems to have less in common with others, corporate-focused parts of the group. Is this still an activity you wish to develop?

Yes, TLS is a fantastic company. Unfortunately, TLS did not win the renewal of a significant contract last year. So it's not the same perimeter. Still with the new perimeter, it's a fantastic company. But clearly, we are clearly -- and systematically, we are clearly open to study what makes the most sense for the shareholder. We have not taken any position or any decision today. But the question is not irrelevant.

Okay. Any additional questions from those in the room?

[indiscernible] journalist and consultant in London. How do you -- this may be a question for Thomas. How do you reassure the workforce and partners around the issue of jobs because that's always the motive issue when we think about AI, and you've articulated really well the implications. But do you get nervousness from TP employees, but employees of your partners and clients as well? And how do you address that over what the implications are for people's jobs and incomes?

Look if you CTP today, we are present in 100 countries, and we have demonstrated growth for the last years. So we talked about the example of India. India has grown now to more than 90,000 people almost doubling the business over the last years. This is the best practice if you're investing in businesses that growing, you see the success on the ground. So TP as a whole has grown its employment numbers over the last years, what the future brings depends on our hard work. And it's too early to say what the net balance would be.

There is also something that you have to keep in mind, which is at Teleperformance typically you have 85% of the workforce on the front line and 15% in management support and administration. Typically, this frontline business is a very demanding business by the way because it's quantitatively and qualitatively very demanding because it's emotionally demanding. And all our industry has a significant churn. Typically, people would say would stay in the front line, something like 2 years or something like that, which means that if any adjustment would be needed at one point or another, it would not present the painful characteristic that is present for many other kind of industries.

Okay. Thank you very much, everybody, for your time, questions. Great dialogue so far today. And for those still here in the room, we'll continue our discussions outside.

Thank you.

Thank you.

Thank you very much.