Endava plc (DAVA) Earnings Call Transcript & Summary

Thanks for joining us here as we are wrapping up the -- about to wrap up. We still have a keynote after this presentation here at the Morgan Stanley TMT Conference for 2026 on Tuesday, so the second of 4 days. Very thankful to the Endava management team for joining us. Before we get started with them, I'm James Faucette, senior IT services analyst here at Morgan Stanley. And we're very pleased today to have co-CEOs of Endava, John Cotterell and Alastair Lukies. We also have Mark Thurston, CFO. Before we get started with the team, though, I do have an important disclosure to read. Please see the Morgan Stanley Research Disclosures website at morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative.

So I guess with that, I'll just kind of open it up. I'm sure you've gotten the same question multiple times as have we. And it starts with, hey, the recent commentary that you gave coming into the calendar year was actually really encouraging. And I'm wondering if you can walk us through the key factors behind that change. And in particular, what are the things you're looking at that inspires confidence in a stronger fourth quarter pipeline conversion rate?

Sure. I mean we've been, as I've articulated fairly frequently over the last year or so, on a strategic focus around, number one, getting into the C-suite and having conversations with our clients about what it is they're trying to achieve with their business and how AI can help drive that. Secondly, around moving more towards outcome-based contracts, aligned with those client objectives that we're talking about. And thirdly, establishing an AI native approach to delivery, which we call Dava.Flow. When you put all of those things together, it's a very powerful proposition to clients. And that's what's been generating the pipeline that is giving us the confidence in the guide that we have put forward. And Al, do you want to just expand a little bit on the sort of client-facing side and how the reactions that they're having to that?

Yes. I mean, I think, we -- a lot of our peer group and a lot of people in the technology world, certainly in this hype. We're in at the moment are selling a fascination or a religion around the technology. We're being very disciplined about talking to the customers, actually listening to the customers, two ears and one mouth, about what it is that they're trying to achieve with their business and then seeing where we can apply AI. So if I look at the shape of the business and the way we're engaging in the C-suite now, recent wins like Paysafe with the CEO, Nexus, the biggest shift in the payments industry probably for 50 years. The rest of the pipeline where what I'm seeing is CEOs saying to CIOs, "If you can go and get me efficiency using AI, knock yourself out, right? You cut my costs. But in terms of ideation and where we can go in terms of competitive advantage, I want to talk to an expert in the field that's doing that." So it's a very different approach to, I think, most companies in the space, and it's starting to resonate.

Got it. So if that's the message, Mark, how do you bring that together and think about what the implied assumptions are of that current pipeline and what portion of it needs to close to get to kind of the targets you've set for the June quarter?

I mean the June quarter in terms of Q4, there is a step-up, which we articulated certainly quarter-on-quarter, 8% on Q3. Part of that is because of the number of days in the quarter. So we get about 2%. So you're looking at 6% uplift. And this is all at the sort of midpoint. It's underpinned contracted and committed. I think we're about 70%, 75%, which is typically normal. It's also underpinned by the recent sort of deals that we have won as well. So we have great confidence in it, but it does also have a wide range on it, something like $10 million, which at this sort of stage in the year, we would usually sort of narrow because experience has told us over the last couple of years that things cannot play out as you might think they do. So it's underpinned by the deals we won earlier in the year, and it's the strength of that pipeline, that conversion assumption.

Got it. So I want to pressure test a few things, starting with kind of your largest clients, and then we'll talk about geographies outside of your top 3 customers. But how have spending intentions been across your top 3 customers? Is this message of the pieces that you can bring together, John, resonating, including with them? And how are they thinking about how much they can you engage with Endava and what the opportunity set is there?

Yes. So that's exactly what we're finding with our largest customers, is that, that level of conversation around what is it you're trying to achieve with the business and us working through how we can bring the appropriate programs, technologies, outcome-based contracts, the Dava.Flow approach to bear, is getting huge interest and huge traction. It's not only our existing customers that we're having those conversations with, but that's where we've had the earliest traction and are seeing the fastest returns out of those conversations.

Are we on track to be able to stabilize the businesses with those largest 3 clients? Or is there still kind of work to be done to flesh that out and to get them kind of on track with some of the newer capabilities of Endava, you think?

So I would say, yes, there's a lot of stabilization that's been done. There's been contracts and extensions and SOWs opened. We covered that on the earnings call. And actually lots of opportunity because the clients are also going through significant change themselves in their businesses and in their marketplaces and in the investments that they're making. And we're seeing opportunities come off the back of that, which we haven't really got in the forecast.

Got it. Interesting. Interesting. And then what about by geography, outside of these top 3 customers. Can you look across your -- where you're operating globally and say, "Oh, this particular geography is strong, this one is weak, this one is so so." Like I'd love to hear about what you're seeing geographically.

I'll let you take.

Well, it's sort of dominated by the end industry vertical. So we're seeing strength somewhat resurgence in financial services. And for us, it's payments, we've just been talking about, but also banking capital markets and insurance. And it's dependent very much on the relative strength of those sectors in each of the sort of geographies. So we're a U.K. heritage headquartered company. We see sort of strength in the U.K. basically. We're seeing also come through in rest of world. I mean we were just referencing the Nexus sort of deal that we announced on the earnings call. So that is causing some uplift in that geography. And also North America, we've -- it's our biggest geography as a percentage of revenue. There's also a lot of momentum there as well.

Got it. So I want to spend some time just talking about the AI initiatives within Endava. Obviously, key and central part of your investment focus and where you're putting resources, becoming more AI native. What are the objective KPIs you guys will use to prove those investments are translating into higher win rates, faster sales cycles, better unit economics, et cetera? Like what are the things that you're tracking and that maybe you can share with us at least from time to time?

I think you just covered a few of them. But yes, I mean, obviously, seeing that come through in revenue growth actually driven by those capabilities. And we're tracking which projects we're using the Dava.Flow on, et cetera, so that we can actually track the impact that it's having. I think margin improvement, so top line growth. Margin improvement, we'll be able to see that on those projects, many of which will be outcome-based that we're actually able to track a margin improvement coming through. So on average, actually lifting the overall margin of the business as we shift to more outcome-based and Dava.Flow-enabled solutions that we're putting in place for the clients.

So -- and those were good financial metrics. What about operational metrics? Is there some sort of productivity metrics or output-related metrics, et cetera, that you're seeing and can talk about?

Yes. Operationally, we've sort of drunk the Kool-Aid ourselves and brought a lot of efficiency to our legal department, to our finance department, to our platforms, to our sales process. So we've reduced the number of salespeople we have substantially and replaced them with higher-quality C-suite engagement salespeople. I think that one of the key KPIs that I would be looking out for is longevity of partnerships. Because I think if you look at Paysafe and Nexus, these are 5- to 7-year partnerships and none of us know where the world is going to be. But if we're still in that dialogue contractually and we're finding efficiencies through Dava.Flow, our ability to create that operational leverage and the unit economics that we're searching for is far greater than getting back into a bidding war every year against competitors who are also desperate to win those accounts. So my job here is to make sure that people that want to come on the journey with us are prepared to put some skin in the game. And back in the day when I was building companies, that was really sort of joint-venture model. So like, are we really going to partner here? We're not setting up loads of joint ventures, but the culture and the cadence of the relationships is much similar to a joint venture.

And how -- like -- and it seems to me from the outside that, that would be conducive to some of this outcome-based pricing and projects, et cetera. Is that right? And like how do you -- and so on the flip side of it, if you're trying to put together outcome-based agreements and as you're saying, Al, your putting -- and that's resulting in some JV, et cetera. But how does that survive? Or how is that reevaluated, particularly if it is a longer duration contract and beyond the horizon of what you could really see?

Look, I think you have to be -- particularly, in a hype cycle, you've got to have some good self-awareness. The history shows us that the projected change is never as quick as people expect, but it's more profound in the long term. And I don't think this is going to be any different. So we've designed the new business structure in a way to say those that are doing more traditional Endava services, which, there's tons of appetite, right? You have got to make sure there's no decline because it's the cash flow from that, that's helping us fund the new. So if a relationship with a strategic partner starts in year 1 with 90% traditional Endava, 10% Dava.Flow, can we get to 50-50 by year 3? So they're the levers that we have to pull to get the blended margin back up to the 20% pluses, into the 30s where we've been before, but it's a blend. And it's -- people keep talking about flicking a switch. It just -- it never happens like that.

Right. So borrowing from some of that consulting developed adoption curves and emotion that goes with that. How do you -- what do you need to do then to -- if we're in the hype cycle, to limit the depths of the valley of despair and as people kind of come to grips with what the realities are versus impressions and then ultimately make that profound change.?

I joined -- just quickly, sorry, and then I'll pass it. I joined because I believe that John pivoted earlier than anyone else. So I think we've been through our trough. I think our trough is done. I actually think the trough of disillusionment, we're now coming back into the plateau and starting to scale. And I think a lot of other people are about to enter it. Go on then.

Go on then.

That's exactly what I was going to say. The -- and if you look at the sweet spot for us, is that combination of we're talking to the CEO, very senior in the C-suite. We're understanding what they're trying to do to their business, what their aspiration is, what the game-changing thing is they're trying to do. We're bringing an output-based outcome-based contract to bear on that so that the client sees we've got skin in the game, we were all pulling in the same direction, but also is giving us opportunity to make wider margins. And then when you're putting the Dava.Flow capability, which is a method, by the way. It's not some platform, we've been using in AI, and I'll come back to that in a moment, if you want. When you get that sweet spot of putting those 3 things together, you actually are giving yourself the opportunity to drive much more significant growth with the client, much higher margins and strong execution. And we have very, very high NPS score as a company. That is because of our execution. And that's why clients trust us and actually go, "Do you know what? I'm going to believe in you guys to actually drive this change for me."

So John, talk about Dava.Flow from your perspective, like what is it? And how does it differentiate what you can deliver to the customer?

So, let me wind back a moment. If you look at the digital wave that we wrote for 20 years, that was largely an Agile-based method, right? And Agile is all about coordinating human beings in the creation of software and doing it in an iterative way. That is completely inappropriate to an agentic AI delivery model. Dava.Flow is the creation of that delivery model that you need for an agentic solution. So upfront, you use agents to help envision what new products or capabilities or strategies a client should be searching around, coming up with options, helping them choose. The next phase is around creating the backlogs, doing the specs, getting the regulations, the governance, the coding standards, everything together that then becomes the ability to do the prompt engineering and the context engineering and all the rest of it that goes into the next phase where the agents actually build the code, governed by people. And then the final phase is the -- you're getting into the support mode of systems that have already been taken live and the improvement of them and so on. Now that is a very different approach to an Agile approach. It's completely different. Culturally, it's different. Agile teams, when they start, they get together. Within 2 weeks, they're kicking code out. The -- in an agentic model, you're not doing that. You're creating an understanding of what you're trying to build and you put weeks into that, and then you have a very fast build and refinement process. So it's culturally different. It's a different conversation with clients, different expectations to manage and so on. And needs codifying it in a lot of detail so that an engineer who is picking something up actually understands the role they're playing in a large program and the phase that they're in and what they need to do. We have captured all of that. And we found our people on very fast learning curves, using the capabilities that we put in place, to actually make it work. We're not seeing other people do that yet.

Right. So can I ask you, like it's a potentially incredibly important point that you're making there. So it seems like most of the metrics I hear thrown around about effective use of agentic AI or that kind of thing, basically just comes down to speed of code production, right? But it sounds like what you're saying is like if that's your focus and your metric, that may be a key risk of pushing you into this trough of disillusionment because you think it's just about speed of production, whereas like it's really the outcome and getting that right from the get-go instead of trying to iterate your way there. Is that fair? Or is that a lacking upfront?

It's absolutely fair. And it's particularly true in an enterprise environment, right? I think people can look at what is happening at a consumer level or an individual who can use these tools to create a lot of code very quickly, the vibe-coding type thing. That does not translate into an enterprise environment easily where you've got regulators, you've got governance, you've got security issues, you've got legacy systems, you've got data problems. All of those things have to be integrated into an architecture and a design that can be fed into the agents so that they code something that works. And it's capturing all of that, that Dava.Flow is all about.

One of my favorite books is Pillars of the Earth by Ken Follett, which goes back to the original cathedral builders. And they used to start laying slabs really quickly, and they kept on collapsing because they didn't design them properly. The job of a modern day an Endava engineer is to do the architectural thinking, think inception, and establish in their mind with the client what is the building going to look like when it's done and not then hand that over to a load of coders, you hand it to the agents to each do their little bit of the LEGO building. So it's a very different approach.

It's an interesting construct that you're building. At least to me, it's somewhat resonant because if I think about the way that things were built for a long time as you just kind of started. If I start with the log cabin, I start -- I see which -- what my trees are, what fits, what fits together and kind of architect as I go. Then to your point, is that, once you start building big things and you've got a lot of labor that can move quickly, you got to figure it out beforehand. And so...

Which AI can help you do, by the way.

Yes, Yes, yes, which is kind of interesting, right? But then if I go back to kind of what people -- like, if I go back to some of the big failures in software initiatives back in -- at the dawn of the Internet ages, people felt like, "Oh, like we fell short of capability and had a lot of cost overruns because we tried to overengineer upfront without really knowing what the potential was." And so that's kind of, in my mind, maybe overly simplistically gave rise to this Agile coding approach, et cetera, where once again, you're kind of back to building as you can, especially in a cloud-based environment, your costs were low if you made a mistake, et cetera. But maybe as you're suggesting is that with agentic development, it seems like the architects start to become a lot more important again. And it seems like what you're saying.

It's not just the architects but I...

Yes. But like the architecture that comes together.

Yes, you have to put the work into what exactly are we building and then you tell the agents to do it. Right. But by the time we're equipped with all of that information, the prompts, the ask but also all the contextual information around this is the environment you're going into, this is the regulator requirements, et cetera, you can get -- and our experience is you get higher quality code than with very good engineers, if you do it right. If you don't do it right, if you're substandard in the way you do it, you have a rubbish in, rubbish out problem.

Got it.

And just to add to that, if you think of our strategy to engage with the C-suite and ideate on what their modern building might look like, what their skyscraper is going to look like, their Salesforce Tower, is if they've already done that work internally and they've got it wrong and it's got to IT procurement and they're just doing an RFP and you're in a race to the bottom against the Indian outsourcers or against one of our peer group, the chances are it's not going to be a great project, and you're going to build a reputation for that. By being in the ideation stage, you can really work through the flow so you get success.

Got it. Any questions from the audience here? Just -- a question here.

How much efficiency will this method bring clients in terms of the job. So if you think about -- there's a lot of software companies talking about engineers doing 20x the amount of work as they used to do per head. If you had a job that was 6 months or 9 months, how much -- and obviously, you guys want to participate in the benefits here, but how much quicker can you -- do you think you can do it?

So we're in multiples rather than percentages, right? We're definitely in that space, i.e. are we talking 3x, 4x, 10x? It's that sort of -- and it depends on the specific job. I think that's actually essential because in order to deliver successful solutions, you've got to do things like address the legacy system problem. That means you've got to lower the cost of fixing those issues to a level where it becomes viable which it hasn't been historically. AI actually enables that, and it's the productivity that enables it. But it unlocks a huge amount of work that's been locked away because no one could afford to do it.

Yes. So it's the obvious follow-up. So if your multiple is more efficient, is there multiples more work there to replace what you were doing before? Just not that in terms of the market size, obviously, you could grow share or whatever, but just in your client base?

I think for an organization like us, there is, right? Because we're shaped right for an AI world with enough senior people and not too many junior people. I don't think you can apply it to the whole market and go everyone, including all these armies of junior people are going to be kept busy by this new method, right? I think that's where you get the winners and losers. But I think for an organization like Endava, with the structure and shape that we've got, there is plenty of opportunity to eat up the productivity and do more with our clients, i.e., they'll still spend the money.

Yes, go ahead.

And can you talk to any pilot or test projects where this has been successful already, maybe on the metrics of what it would have looked like versus what it looks like now?

So yes, we're not actually disclosing any of that at the moment, right? Some of these are 12-, 18-month programs, and we're seeing the benefit, but we're not actually able to reach the point where we have completed and actually able to demonstrate that the outcomes that we got were what we expected or better. So we're seeing it, but we haven't reached the point where we can really produce definitive things on it.

Let me just add. Because, obviously, under NDA with particular customers, we can demonstrate how the methodology has been viewed, because we had to show examples. So the Nexus bid that we won against pretty much every big IT services company in the world. So it's us and AWS that won the rebuilding of SWIFT, if you want to think of it like that, starting with 6 countries in Asia. It's an enormous project. I mean it's a huge project. But it's under some time pressure, which has now been compounded by the geopolitical environment because it's all about sovereignty of data and do we want 2 big American companies managing these systems around the world. And so we're on an aggressive time line to get that delivered. They would never -- if you look at who Nexus is, set up by the Bank of International Settlements as a joint venture between some pretty big players, they would never have picked us if they didn't have the confidence that our methodology would get it there in time for a modern platform. So a lot of the deals we'll be announcing and the deals that we've announced are based on us exposing the methodology to people. We do get asked the obvious question, which is, 'Does that mean we get it for less money?" No. But it's a much more guaranteed outcome, and it will be more efficient when we do it. So there is an element of alchemy. We're not a company that builds IP per se. We don't build our own platforms. We're not committing to platforms. But we do think the methodology, just like the Agile Manifesto did, will become something that people gather around. That's our ambition.

And just you reminded me there, the thing about Dava.Flow is it's tool agnostic. So clients are no longer worrying about, "I have to decide whether I'm going to choose Anthropic or ChatGPT," and "Someone else will come up with something better and my whole choice has been thrown out," and "I don't want to start a project because something keeps coming over the horizon." Dava.Flow takes that away because you can plug whatever tool you want into it and then execute against that.

So last question here in the last 1.5 minutes, and I want to take this back a little bit to the P&L. How much of recent margin pressure is intentional, that is investment in this and other transformative initiatives versus structural pricing utilization, going back to kind of this question here? And help us understand the impact -- relative impact of margins and when we should expect to turn.

I'll let Mark give some definitive numbers. But essentially, we've been investing significantly in the pivot, by investing in the AI capability. We've invested in staff with AI capability that are not fully billable yet. And so that's had a depression impact. And our billability has come down because we're not driving the growth rates that gives the healthiest level of billability right now. I don't know whether you want to put number on it.

Took the words out of my mouth. I mean, quantification, it's about 3% that we've been investing. It will abate as we go through into our next fiscal '27, but we will continue investing but not at that level because the groundwork has been done basically.

Great. Well, we're out of time here. John, Al, Mark, thank you very much for joining us here at the Morgan Stanley TMT Conference, fascinating conversation and best of luck.

Thank you, and a great conference. Cheers.

Thank you.