All right, we're going to be kicking it off here now. We have Endava, a digital engineering provider with over 11,000 global professionals, heritage strength in payments, and a base of operations remains Central Europe, but you've been growing globally, bigger base in India and APAC. With us today, we have Mark Thurston, Endava CFO since 2015. Mark, first off, thanks for being here. Not at all. Good to see you. We're going to be getting into GenAI, Dava.Flow, your partnership efforts. Before we get into that, let's start on demand. It's still fresh. You had earnings just a week ago. I wanted to start by talking about some of the changes, the client decisioning, the demand that you've seen. Really just talk about what the latest is in that enterprise conversation that Endava's having. As you can see, it's been volatile. We missed on the top line. We're making a pivot anyway at the moment, where we're trying to go for longer-term, larger engagements, more where we're acting as a partner with the individual sort of clients. They're big moves for the clients as well as ourselves, and timing of those because of due diligence processes is uncertain. I'd say demand, despite the miss, seems to be firming from an AI perspective. It's moving out of proof of concept to actually trying to put into production. We're seeing a lot more of those conversations in the pipeline, something more meaningful. It's just predicting them is very difficult, and I think that's, particularly with us, given our sort of scale, we tend to be at the leading edge in terms of technology change as being problematic. We have had things such as the Middle East, which impacted us with clients directly in the region. We also felt a little bit of a chill, I think, from the conflict in terms of uncertainty amongst the client base, particularly banking and capital markets. Okay. It's a mixed picture, as you can tell. Just on that Middle East point, just what is that direct exposure and how broad was the pressure in that base of clients? On the indirect side, how broad-based was that chill and that thought? Yeah. Just putting it in numbers, I think we reduced the Q4 guide by about GBP 15 million from GBP 200 million odd, and GBP 3 million of that was clients that we expected to deliver revenue in Q4. Okay. They were mixed. They were shipping predominantly sort of payments and banking, so a range, but they basically just stopped. The chill effect was mainly in banking and capital markets, I'd say evenly sort of split between North America and the U.K. A pullback. It's very difficult to, causality, what event does it lead to behaviors? That was our sort of general sense of it. Okay. Aside from the conflict in Iran, there's different streams of activity and work going on. Can you just kind of go back and detail the trends that you are seeing, even just before that conflict, as far as the heritage discretionary kind of digital transformation work that Endava's been known for? Yeah versus kind of the now newer AI-related activity? I think, I don't really want to use the word traditional, but I will. I'd use legacy. Legacy. Heritage. It's going to be there for a long time. Right. Things don't change that sort of quickly. I'd say it's been under pressure, and I'd say the time and material model particularly. Our day rates in terms of the average bill rate per head or work day has remained sort of steady through this sort of period. It's been volume that's been the issue. Okay. I think it's this transition to the new AI world. We've seen the legacy traditional business sort of under pressure while people are thinking about what the real digital shift is, going from what was traditionally digital to the AI sort of shift. We've been thinking about that in terms of Dava.Flow, which is our approach, our delivery approach to AI using agents to deliver sort of work. It's also then how do you change the commercial arrangements around it? Move away from the traditional time and material to more, there's a word used, outcome. Let's call it fixed price. There are varieties of, let's call it consideration you can get within that sort of framework. How do you make that sort of shift? I think this time last year, sort of June 2025, we were about, I think, 77% sort of time and material, and the rest was fixed price, traditional sort of fixed price. Within there, you would've had relationships like Mastercard, which are a sort of flat rate within that fixed price envelope. It's moved as we've been trying to move the business model to something like 73%. There's been us in terms of winning and wanting to win longer, larger term contracts because they give us visibility. Right. They are also, if they are priced and the commercial arrangement is on an outcome-based or fixed price basis, that the economic benefit you get from deploying Dava.Flow, you capture more of that margins through that sort of process. They are more difficult things to negotiate than a strict sort of time and material sort of contract. I think that's the inflection point that we're going through. We're seeing positive movement. Obviously, on the call last week, we were saying 15% of our revenues we're saying are AI native. They are a combination of these largest deals, such as a Paysafe, where we intend to deliver through Dava.Flow. Also the relationship also with the hyperscale. The business model It's pivoting. It's another word I don't quite like, but it's moving slowly, but it's moving in that direction where we'll be, I think, more of an outcome fixed price business. There'll be more of the work delivered through agents, Dava.Flow being the methodology around it. It'd be better margin. Yeah. That is the pivot that we're going through. It does require clients to lean in, because some clients don't want to. They want to contract on the traditional sort of basis. It does need forward-leaning clients to engage with as well. It's happening during a time that seems like just ever-increasing geopolitical volatility. You can understand why it's choppy right there. Yeah. As it relates to that, as far as the chill behavior on some clients, your June year-end gives you six months right into the calendar year here. As you talk with clients and you consider what they're thinking about in their calendar year budgets. Yeah Are they cutting things or is it just kind of a wait and defer kind of behavior that you're seeing? We're not seeing cutting. We see project work come to an end with some clients. Got it. We tend to be project sort of driven, which is part of the visibility sort of piece. It doesn't feel like they are stopping things. They're just moving very sort of slowly. We've been having conversations with some large clients, it feels like an eternity. Six, nine months, we thought stuff would be done this side of the financial year, and it is continuing as they go through that sort of process. I don't think budgets are being pulled back. We do sense there's a sort of change in terms of AI adoption being more enterprise-wide, what is the next sort of step for doing that. I think budgets, we're not seeing a cut. I think the money is there, it's just how quickly it comes through. Some slower behavior for some clients, an aversion to change model, as I imagine procurement has to get comfortable with all this stuff. Yes. Yeah. On top of that too, a shift of more large deal-. Yes focus at the same time. The procurement thing is very relevant because with time and material, you have a lot of visibility. Yeah in terms of the cost and where people are located, and rate cards, you can compare against another vendor. It's to take that discussion to a higher level, more strategic, more C-suite, where it's, again, going back to that buzzword, outcome-based, more visionary, longer term sort of partnership. Trying to circumvent, in some respects, some of those discussions you have with procurement and take it to a sort of higher level. Yeah. Okay. Even in the agentic world, you have companies right now learning the hard way as far as- Yes scoping costs and everything else blowing out. Yeah. There's a lot of moving parts, clearly. Yeah. Okay. On the large deal front, can you talk about that? That's something that historically the company, the heritage digital transformation work would be kind of smaller scale, felt like it would build in programs. Yes. it seems like you're going after broader- Yes end-to-end work. Is it a function of a conscious decision by Endava? Is it something in the market that changed in the last couple of years? It's certainly something conscious with us. We've had large engagements with the likes of Mastercard and Worldpay. They've sort of developed from small seeds. Right They've sort of scaled, and the longevity of the relationship has resulted in a longer term, multi-year sort of commitment. Right. Now we are tending to go for it from the get-go, sometimes with clients who don't necessarily know about us. The pitch is higher. It's not an engineering technical level. It's about going back to the outcomes rather than being very technical in terms of what we can do. It's what outcome are you looking for? What AI adoption level do you want to achieve with the company? It's a very different pitch. Right We're getting our muscle around that as well. You need industry knowledge, which we have in spades in the payments and financial services space. It's bringing in others who can follow through in sort of TMT, et cetera, where you can get that strategic engagement. The go-to market has changed. With larger deals, they're binary. They're a more expensive chase. Right. They can be longer term in nature and a little bit more uncertain. My sense is that might be a transitory phase we're in at the moment. We do feel there's a tipping point in terms of enterprise-wide adoption of AI, given we were just talking about the economic sort of background uncertainty. I think we are getting through that sort of transitory phase. It will be a little bit more lumpy. Okay. As far as in your transition to pursuing these and you think about pipeline composition, large deals as a mix of pipeline or how that pipeline looks. Can you comment on that? Yeah. Despite the results last week, it is strong and conversations are happening. They don't proceed at the speed that we would like them to. Yeah. We delayed last week's call by about a week because we thought we would have some good news to put out alongside a poor set of figures. Nothing to do with the client not wanting to move at speed. They do. There's physically a lot to get through to get you into the stage where you can contract and sort of commit. We are finding that with most of these engagements as well. Some of the engagements are not straightforward. They can be efficiency. Take the IT function, whether it's developing or managed service or whatever, from a client and run it more efficiently. Others can be build. We talked on the call about PGx, which is our pre-built payments gateway- Yeah or minimum viable sort of product. That is going to be more sort of lumpy in terms of the deliverables, because we'd be dropping code. It'll look a little bit more like software to the client as we sort of have pre-built it. Those conversations where it's about our product, is about can you deliver it at a cheaper price more quickly? It's also, can you deliver other benefits to us? Cost reduction being one of them, but can you attract more volume to the platform? Okay. We will take some gain share in that. There are benchmarking conversations that go on as well. They are not straightforward arrangements to negotiate in detail legally, which is also part of the elongated- Okay sort of timescales. Okay. Now as you think about building this muscle, obviously a lot of volatility in the backdrop. As you think about the guidance framework and going forward, what actions are you taking to try and minimize surprises? We're getting more granular, certainly, in terms of what we look at in terms of the clients and the opportunities. There's always going to be uncertainty, and I think it's heightened now. Because of the size of some of these deals, they're not immense, but they're chunky. If they move out a quarter, it certainly hurts a company of our scale- Right quite sort of significantly. Again, it's trying to use judgment. We use probabilities about where opportunities are in the pipeline stages. We try and make those judgments not on a monthly basis, but if they're significant, we want feedback on a almost daily, weekly basis about how things progress. I think it's further scrutiny of those probabilities. I think we will inevitably get things wrong. I think it's also bringing in more mature managers actually into the business, who can actually make those judgments for us rather than it being done top-down. Okay. more rigor, more granularity, and you carry that forward here. I assume 4Q- Yeah any large deal conversion. Yeah Probably. Okay. Yeah. Okay. Let's shift down to AI and maybe the client adoption. You mentioned this a bit before, but just how has this, as far as GenAI and agentic adoption in your bigger clients, how has that evolved? The pitch, if we go back to this sort of AI native business model, we were talking about the 15%. Yeah that will include the likes of Paysafe, RELX, et cetera. Those deals were basically won on us committing to a multi-year relationship with them and delivering the benefits to the client through AI. AI is embedded in those deals, basically. It's coming, but they're not felt immediately because usually there's an onboarding under traditional methods, and you're planning for the deployment of Dava.Flow. Those engagements will also be a mixture of T&M, fixed price, as the various packages change over the duration of the contractual space. There's planning to push Dava.Flow in. It is in place in some of those engagements that we have. On the other side, we will have smaller engagement with clients who want to move quickly and want to experiment with what Dava.Flow can do. They typically, in those circumstances, are, let's call it creating a lab to experiment with Endava in terms of what it can do. The next stage is then how do you roll that out across the organization. Which is what we're doing with the bigger organizations, where we're actually taking things lock, stock, and barrel, and then evolving it as part of that contractual arrangement. Some of them are small experimentation in the client that we hope will evolve into something bigger. Okay. Well, I guess as far as this approach goes, what do you perceive it to be as far as the differentiation versus competitors and what's out there in the market? Talk a little bit further about Dava.Flow and exactly what's different. Yeah in that approach. Dava.Flow is similar-ish to distributed agile, which is our methodology as a team is where it was, we called it ideation to production. It's a proof of concept pilot, and then it scales people involved. It's a similar sort of process because people are involved in Dava.Flow. We call it Signal, Explore, what's it called? Govern, and then Evolve. It's that similar sort of thing. The Signal stage is sometimes can start in the sales process, where you're taking on board inputs from the client and coming up with numerous options very quickly, and that would've been a proof of concept. You would've been physically listening to the client, interviewing them, looking at architecture maps, trying out one sort of concept with the client, and then going back and forth. With Signal, it's also part of the selling process. Okay. You can talk to the client about what their strategic issues are, but you will have, I don't know, you could have 10 variants about what the solution may look like. When it moves into Explore, you move very quickly into an actionable backlog. There's not this time lag where you go through the discovery phase, and then you need to go through a validation phase physically. It is able to come up with something that is sitting in the backlog. It's almost ready to go. You just need to decide what to do. Getting back to what it sort of is, it's again, a methodology. It uses agents to do this stuff. We have partners with OpenAI, we can use any large language model because the client may have a preference. Right. They may have a preference for where they want it hosted. There are various tools we use, Miro and Cognition as part of Dava.Flow, you can use any sort of tool depending on what the client wants. You're sort of tool-agnostic, although we have preferences. It's the methodology of doing it, and basically, it's the governance and control on the quality of the whole process of agents working alongside humans. What we've seen of it moves very fast, and sometimes the client is the bottleneck. We are producing so much stuff that they are the bottleneck in it actually accelerating further. This is even before we've got to the Govern phase, which is where the backlog is released, and you're writing code and you're producing product. We're in that very early stages of Explore and Signal. I don't think we're yet in Govern, which is where the models are producing, the agents are producing the code and the product, et cetera. Okay. Now when people have seen it, and we have various diagrams of how it conceptually sort of works in our investor deck. They are very excited because it is different to what is being sold or pushed. People talk about platforms, so you plug into our AI platform, and it will do all of this for you. Where we do not have a platform. We have a methodology using tools and approach and governance around it. It is very adjacent, I would say, to the teams methodology. Methodology using AI and agents, and we have a methodology for T&M because the two have got to coexist- Yeah for a very long time, I think, certainly another three, four, five years. It sounds flexible, so that it should help you or help clients adopt this. Yes Certainly speeding up the pace of effectively revenue conversion. Yes is what I'm hearing. Yes versus the traditional model. Well, the key thing is how if it's outcome-based, you can get there a lot quicker. Right, and stickier. Yeah. Say if we have a build process, which we are in one instance, we're building something over 18 months. It's based on deliverables. Right Producing stuff. We've costed it and done those timescales around the traditional T&M methodology. If Dava.Flow comes onto it can, in theory, accelerate those milestones. It will come down to whether the client is ready. Right for the deliverables when they want them. You've gone, I think, three clients two quarters ago to 12 clients this quarter. Yes. How quickly can you get the broader client base on board? I think there's an element of the sales force needs to be properly trained in it. Yep. Our delivery staff, we're investing a lot in training them on it. I think we've got 1,000, I call them practitioners in Dava.Flow out of 11,000. They need to be trained in it. There's getting it ready beyond the, let's call it the cottage industry stage, up to let's call it production, where we can push it at clients. It has to be done sensibly because we don't want to cannibalize revenues. The client has to be ready for it and engaged in it. Some clients don't want AI. Sounds a bit absurd, but they don't sort of trust the controls and the governance around it. You'll have procurement functions that don't want it because it's complicated, and it's an outcome, and now on earth am I going to sort of change this? I prefer through time and material. It's going to be a little bit baby steps with the existing sort of client base. Okay. I think where it gets the step change is the new deals, longer-term deals, where the C-suite wants a five-year vision to transform their company. They want Endava as the partner to enable them to do that. They want to learn about Dava.Flow as they go because they're experimenting- Right with these tools, and the approach resonates, and it's who's going to sort of guide them through this process so that it achieves the benefits they want and it's safe at the end of the day. Understanding it's early, how does this impact those engagements? For those clients that are a little bit further down the road in leveraging Dava.Flow, what does that do to the revenue model for you? What does it do to the margin profile of those engagements? Talk about that transition. Dava.Flow predominantly will be, let's call it fixed price outcome-based. In some respects, it could be lumpy. If we're, for instance, dropping a PGx, it's almost like 100% profit, but we'll be plugging it in. We are also working on connectors as well, which are pre-built. We're sort of blurring a little bit between having a modular approach instead of where we were in the past. We'll build from scratch again. Okay. We're only doing this at the moment, basically in the sort of payment space, because we know it well and we know the solutions, but it has a lot of applicability to other sectors. One of the PGx drops was to a utility company, basically. Sure. It will have a characteristic of being sometimes a bit lumpy. We do a drop, it's very profitable at the edges. This will be like GBP 5 million or whatever a quarter, but it will be longer term, steady revenue, basically. We may have, over time, going back to the outcome-based where there is revenue share, where the platform performs, we'll disclose this because it's, in theory, revenue not under our control that we're taking a share in. We'll be very clear about that. You may get some of that- Right regulation. It should be steadier, higher visibility work, a little bit of lumpiness where, I'll call it product or there is a drop. Maybe a little bit of lumpiness where there is revenue share in that sort of construct. If we get that right, the profitability on those contracts will be a lot higher than the traditional time and material. Typically, on time and material, we're about high 20s. Profitability on these projects will be well in excess of that. Okay. as you think about longer term profitability and ability to potentially get back- Yes to where you were. Yes Maybe more. Okay. There's a lot going on, obviously, within the agentic world and GenAI delivery into enterprises. The last couple of weeks, there's been a lot of news flow around OpenAI, Anthropic, through joint ventures, through some M&A- Yep leaning into their own services activities. Yeah. What's Endava's perspective there as far as the competitive threat versus just co-opetition in the market? How do you see that unfolding? I don't see it as a challenge. I think what we have as a services business, we have deep domain knowledge. We know legacy. If you think about them, they're a very new technology and how do they push it into the real world? It's a bit like software and ERP, et cetera. They needed companies like Endava to implement it. There's also been a sort of ebb and flow where product companies have had, let's call it professional services, to do this, or do they leave it to service companies to do? I think as long as you've got good technical understanding, you've got Dava.Flow, and you've got industry sort of knowledge, you'll be well-placed to deal with that. Okay. At the same time, you have a relationship with OpenAI? Yeah Separately, right? Talk about that approach as well and that go to market. We generally do it jointly. They will point us at opportunities where they need a, I won't call it an implementation partner, but let's use that sort of word. They point us in that direction. We get a lot of previews of what is in the pipe, what they are developing product-wise. Also, how they are approaching the market and whether it's a threat or it's going to be a tailwind for us. There's a lot of collaboration between the two sort of parties. They help us internally in terms of our AI native adpption to help us in the sort of rollout and training as well. There's a lot of synergy between the two firms. Okay. All right. I know we're out of time. Appreciate the perspective, Mark. We covered a lot there as related to the GenAI, the adoption within the client base with Dava.Flow. Obviously, some changes in the forecasting going forward amid the volatility. As always, appreciate your time- No sure. appreciate you coming to the TD Cowen Conference. Thank ypu. Not at all. Thank you. Appreciate it.