All right. We'll go ahead and get started. Thank you to everybody here for joining us, and those joining us via webcast.
Early part of the day here on day two of the 2025 Morgan Stanley TMT Conference. Very pleased today to have Endava, John, and Mark, CEO and CFO. Before we get started with that, I do have a quick disclaimer I need to read. For important disclosures, please see the Morgan Stanley Research Disclosure website at morganstanley.com/researchdisclosure. As I alluded to, I'm pleased to be hosting this morning John Cotterell, CEO of Endava, as well as Mark Thurston, CFO. Maybe, John, love to get from you and Mark a little bit of a quick overview on Endava, perhaps for those that aren't as familiar, just where—what part of the services ecosystem you address, but how are you thinking about your business in particular and your fiscal year 2025 outlook?
And kinda help us, c—I guess, calibrate how you're thinking about what can happen between now and the end of your fiscal year.
Sure. Endava is a digital transformation business. We've had many years of growth driven by the digital wave, came out of the dot-com boom, dot-com bust, and led into, you know, 20-plus years of regular, year-on-year growth. You know, that was a period where we largely solved the engineering problems that the internet brought along by saying, "We will build capability around the outside of clients' core systems." You built digital products that enabled them to go to market through multiple channels, and touched their core systems as little as possible. We are now going through what we call the digital shift.
Right.
AI-driven. The difference about AI to how the last 25 years looked is that AI needs to get into the core. It needs to get at the data and the processes that exist within a client's core systems. That's a massive change. It's a massive change of engineering challenge.
Right.
You know, the moment you start going into the core, you start unraveling the spaghetti, the architectures, etc., that clients have avoided touching through all of this period. The good news is that AI is also bringing capabilities that help us to X-ray the core, if you like, understand what our clients' existing systems do and come up with much better pathways to actually opening them up to digital product, AI-enabled digital product.
Right.
For us, it's a period of shift, where a lot of the new programs that we're engaging with have got a much longer sales cycle, because the complexity of what the client's taking on board has gone up a level as you—as you're going deeper into their organization.
Right.
And, they're also, you know, coming off the back of COVID, they've raised the bar from financial officers' perspectives of what a business case should do. So you have technical uncertainty, higher bar, longer sales cycles, and we're going through that period where what used to go from ideation to production in three to six months.
Got it.
Is now one to two years. You are seeing an impact on revenue as we go through that.
And so.
Hang on.
Yeah, yeah.
Let me just give it to Mark.
Yeah, yeah.
Yeah, yeah.
Yeah, you said a lot of so many provocative things I wanted to get to some of them, but yeah, let's get to Mark.
That's being sort of reflected in our revised sort of outlook.
Mm-hmm.
We took the guide down modestly, 'cause we have seen this elongation. More specifically, as we've come into this calendar year, we've seen geographic slowdown, basically U.K. in particular, rest of the world, but mainly in sort of Asia-Pacific. There's sort of, you know, there's sort of specific sort of fiscal issues in the U.K. that have caused caution in terms of the outlook in terms of spend. It's not particularly sector sort of driven. The recovery that we were seeing, which we had initially at the start of the fiscal year of a gradual sequential uptick, is flattened, but it will pick up in Q4.
And, you know, putting alongside what John just said about sort of like pipeline conversion patterns, we see that that will happen based on, you know, the experience of mainly in the digital transformation side of the business, is a characterization of it. We see that that will come through, and we will see that sort of uptick.
Got it. One of the things I was gonna ask in terms of that process that your customers are going through, which I think makes a lot of sense, but a key question that we're asking as Analysts and Investors is, what does the return profile look like on the AI projects? Clearly, like, they can be more comprehensive and, as you said, you need to really get at the data, you need to get at the processes, etc. One of the things that at least we're grappling with is, what do the return profiles of these AI-related projects look like once they're implemented and that part of the equation to help understand what's happening in demand. Any early learnings there?
I mean, that really is a how long is a piece of string question.
Right, right.
There are some areas where you get much higher returns.
Okay.
That, and so, you know, the close—the closer you get to the core use case capability, i.e., chat, the higher and the quicker the returns that you can get, subject to the engineering challenges of how far you have to go into the core. The more.
Hold on just a second. Is your mic not?
I'm getting some interference from the, yes? I'm having interference going on here.
Yeah. I wanna make sure we get that for the webcast. Thank you.
I'm gonna give you this.
Okay.
Let me get you one. I apologize.
Thank you.
Let me carry on with this. Is that working?
Yeah. That's great. Yep.
Yes, depending on the use case, you can see much quicker returns. If you go, for instance, today into a call center environment.
Mm-hmm.
The business cases are struggling to stand up in many cases.
Right.
Right? Because of the processing costs, the token costs that you get on AI, it's actually cheaper to operate it with people today.
Right, right.
Than it is to switch to a model. Now, you can start putting engineering in place around, making some of that run more efficiently, only using tokens where you need to as opposed to universally. It's still struggling to get a case. Now, we've had DeepSeek come along recently.
Right, right.
You know, that we can drive a whole drop in the level of processing costs. Hopefully, that'll come through in seeing that happen more widely on tokens. That will then start helping those business cases to stack up. It is a fast-moving environment.
Right.
As to which spaces you're gonna get the, the business cases to work quickest, and this is part of the reason for the longer sales cycle that we're talking about.
Right. No, understandably. Understandably. And what—and then as you go through that longer sales cycle and you get stuff into backlog, is there any change, though, in backlog to revenue conversion cycle time, or is that pretty consistent? It sounds like just getting the sales process, getting it from pipeline to backlog is extended. What about from actual backlog to con revenue?
I mean, as going back to what John was saying earlier, certainly in the, let's call it the core modernization space, that burning down of the backlog.
Mm-hmm.
Is slower than the initial velocity that we've seen in, call it traditional digital transformation. Now, we haven't got there yet to get those sort of data points, but certainly the initial view that we have is it's a slower build.
Mm-hmm.
To meaningful revenues.
Got it. But it's safe to say, though, especially based on your comment just a moment ago, Mark, is that that slower build, you do have the visibility that that'll start to build in the fiscal fourth quarter.
Yeah.
Right.
Yeah, yeah. That's very near term.
Right. Got it. Got it.
Just to be clear, the ideation to production process, you've clearly got a sales process to kick off the ideation.
Mm-hmm. Mm-hmm.
Through ideation, proof of concepts and so on to actually ramping builds that lead to significant revenue, that's a three to six-month process.
Right.
Now we're seeing that take a much longer process.
Right.
The amount of architectural work, design work, and so on that you need to do as you're analyzing and understanding the client's core, and what you need to change there to enable new product, makes that a longer cycle. It is a longer sales cycle, and it is a longer cycle to get from ideation to production.
Got it. Got it.
Once you're in that second part of the cycle, that's a piece of business that we're working on. We can see the route to revenue, and that's partly what Mark's referring to.
If you have these longer sales cycles, etc., is the relative size of the project changing? Once you go through the longer sales cycle, you have a more complete view of what you need to do. Now you're underway with the project. Are those longer or at least larger? Kinda how should we think about that?
yes.
Okay.
I mean, put simply, larger and more complex, but with a longer route to getting there.
Right. So really what you're talking about here is still obviously open questions about, as you said, it seems to be pretty dynamic where returns might be on some of these AI-related projects. Or if we look at the kind of work you're doing generally, there's more upfront, but once you do, you should have better visibility on the projects themselves. Is that fair?
Yes, absolutely.
Okay. Let's talk about the U.K. Mark, you mentioned that was one of the weak geographies for you here at the beginning of the calendar year. What are you—how are you thinking about the evolution of U.K. client spending intentions over the last couple of quarters? Give a little more color on what's happening there and what are the things that you're looking for that would be indicators that that's about to recover?
The U.K.'s still going to grow, we believe.
Okay.
It's just not at the rate anticipated. What we sort of saw in January, we saw some clients pull back, ramp downs that were unexpected. When you probe into why that is, it usually comes up budget.
Right.
Actually, what underpins that, I think, is the uncertainty of the U.K. environment. I mean, there is a new administration in the U.K. They have been plugging a fiscal hole. The burden has fallen literally on corporates in terms of employee costs. That has created uncertainty. Certainly, the view at the moment is they will not have filled all of that hole. There is an anticipation there will be further sort of tax rises, mainly in the corporate area, that will be coming. The whole scenario looks pessimistic compared to how it was, certainly sort of back in November. I think that is giving clients and CFOs caution about the spend and velocity. I think we will still see growth.
Right.
Not at the rate that we're anticipating, say, November, September time.
Right. That's really helpful color. With those factors having dampened demand, what do you look for on the opposite side as, you know, would help provide relief and make you start to feel better, that the U.K. can start to accelerate its growth further?
Speed of decision-making, basically.
Yeah.
I mean, you've—I think you've heard it from us continually about slow decision sort of making. An elongation in those decision-making processes is, do we see any detection that they are speeding up, that they are progressing as anticipated? We're still seeing more can kicking down the road, basically.
Right. On specific verticals, are you still seeing a mixed picture in the Payments vertical, and within payments, can you help us break down pockets of weakness and strength? What kinds of areas or projects are you seeing strength in the Payments vertical?
Meaning this, this dominated by two large clients.
Mm-hmm.
Largest client, I'd say, is stable at the moment. I've seen big headwinds over the course of 18 months, with a mainly around sort of product suite.
Okay.
It is stable basically in terms of the outlook. Second biggest client is more positive. Certainly the trajectory is they're being increasing spend quarter on quarter. It's a little bit subject again to that word budgets.
Right.
Q4 looks a little bit slower, certainly from them. I think our third largest client, which is a U.S.-listed payments client, has been reducing spend. It's a little bit mixed, to be fair. Overall, I'd say it was stable given those sort of puts and takes. There's no geographical sort of color to it either. I mean, payments, outside of the payments processes, which is where, you know, we disclose the revenue in sort of, and so we'd include work that we do in the payments space in Banking and Capital Markets, is more encouraging. There's also, you know, payments used in other industry verticals that Endava, you know, discloses where there is still appetite.
Got it.
Yeah.
Just to ask you that.
Yeah.
If you look at the Banking and Capital Markets, which has been picking up reasonably strongly, some of that is payments-driven.
Mm-hmm.
You're seeing a bit of a shift in investment from the payment processes to the banks, who essentially are doing the same thing.
Right, right, right.
They've just gone into their payment processing.
Right, right.
Actually we're seeing, you know, customer-oriented investments by the banks who are seeing some of the value out of their payment processing in terms of customer insights and capabilities. There's a little bit of a shift going on in the market that means you'll see, I think, you know, the payments processing side not as strong as the Banking and Capital Markets side over the next few years.
Oh, interesting. Is that shift that you're talking about in banks and bank spend, John, is that globally or is it within specific geographies, and how are you feeling about what's catalyzing that?
I mean, you're probably close to the numbers, but my instinct is it's pretty global.
Yeah.
That it's picking up.
Yeah.
Yes. Yeah.
Got it. Got it. Got it. And wanted to ask on, on pricing. One of the things that we've heard kind of around the, the services space is that pricing's not super aggressive or the competition for pricing, and there's not a ton of pricing pressure, but there is some, and that it's, when you combine that with wage rates, etc., is that it's forestalling some of the margin targets that a lot of people would get to. But really it's a—it's a function of like what's happening from a pricing perspective, and it maybe just goes back to, to customers being very cognizant of their—of their budgets and making sure that they're getting all that they can. Is that consistent with what you're seeing or what are you seeing, you know, more generally from pricing perspective?
I think the last 80 minutes, we've seen pricing sort of stable, which is an, and when we make that comment, it's an, it's an average. We look at the average.
For sure. Yeah.
Per workday. You see there are obviously puts and takes in that. We have been sort of stable. Others have called out pricing pressure. Recently we have seen a slight improvement in pricing, actually, quarter on quarter, which is encouraging. I think that reflects also what others are saying. One of the drivers for that, I mean, I think we are managing to get through inflationary pressures in terms of wage. It is probably also due to the mix of the work that we are doing, which is more highly value-added, such as [crosstalk], etc.
Right.
It is becoming more benign, the pricing environment. John, any comments?
Yeah. I mean, I think the thing that I'd add to that is, the part of the shift that we're going through is the deployment of more of our own IP.
Mm-hmm.
Into the way in which we operate. Essentially, taking AI, and applying it in a way that allows Endava to operate more efficiently, with better assurance of outcome, things like our agentic AI solution.
Right.
which is also applied to our core modernization approach. These are actually significant added value in terms of a client being able to get from us, a higher quality outcome that costs less than if they bought lots of people doing it manually.
Sure.
Which would be in traditional ways of doing things. That is helping us to protect our pricing environment, and actually you've seen the margins beginning to pick up for us.
Right.
In the guide and, you know, in what we've been delivering. Actually, we're beginning to move beyond that short-term pricing pressure, by deploying IP into the way in which we operate.
That's actually another really provocative statement there because I think the over the—
I'm intrigued by provocative.
Yeah. No, I just mean by, by from the standpoint that the dream in services has always been IP development and retention to ultimately get leverage on the delivery, right? It seems like what you're saying right now—one of the challenges is always that retention of IP, particularly if and as it's developed while under contract with a customer. It seems like you're making a little bit of progress there. Can you talk about how you think about that over the medium to long run? Does it really have the potential to move the profitability and type of business that you're running?
I mean, this is what we're very focused on as a business.
Okay.
I think it's something that AI is opening up in ways that wasn't possible before. And just to be clear, we're not—we're not talking about harvesting what we see as client IP.
Clearly.
Right? The IP that they need to run their business. We will build solutions for clients and they will own the IP on that. We're talking about IP that enables us to operate more efficiently, and, you know, to use an agentic AI solution to create something for a client that has a much higher confidence of delivery. Going into their core systems, actually using AI to read the code base and understand what their existing systems do. That is not a simple thing, by the way.
No, but no.
and then being able to use that to establish much greater roadmaps to, pulling out the elements that need to be changed, without breaking all the connections that they have built in, to, other parts of the system, to enable AI products to be created that go deep enough into the organization, you know, to be able to answer questions. If someone calls a call center, you know, "I'm unhappy with my bill," AI needs to understand how the bill was created in order to be able to understand and resolve the question. That's why you have to go into the core.
We can use AI to tackle those challenges much more effectively, than, than you could traditionally.
Got it.
The traditional approach was put hundreds of cheap developers from the cheapest country you can find them in.
Sure.
To read all the code and try and work out what it does. That does not give a high-quality, understanding of what our clients' existing systems do. Whereas using AI does it in a much more structured way.
Right.
You can get a much higher quality understanding of what it's doing. We're focusing our energies on creating that sort of IP. It's essential to what we do, right? Because we're still doing digital transformation for clients.
Right.
It's just that with AI as part of additional products, it has to go much deeper into the organization. Creating IP that enables our people to operate more effectively is a key part of the next generation. It's also an area where I think, you know, if you look at Endava and our history around digital engineering, we are well set up to be able to create those capabilities, and to do it faster and more effectively than most of our larger peers out there.
Got it. Let me ask quickly, you know, just from a demand perspective, a little bit of a left-field question maybe. That is that, there's obviously lots of concerns around tariffs and trade protectionism, et cetera. What's the early reaction that you're seeing from company and company managements in terms of what they need to do, potential impact on input costs like hardware and, and, you know, whether that compute, networking, et cetera.? Any early reactions that you're hearing from your management, managements and customer sets?
Yeah. I mean, this is a very uncertain area. I'm sure everyone is going, all of our guides and everything do not include what might happen over there, because it's unclear exactly what might get tariffed, and exactly the way in which those tariffs would be implemented. At the moment, you know, it looks like services is a space that would not get tariffed.
Mm-hmm. Mm-hmm.
Having said that, Macron raised the concept of tariffing a service in response to something that the U.S. might do. That is the first time we have seen that in a way that might impact our business. That would be the biggest impact if that sort of a tariff was raised. I think that is unlikely. I think across the rest of the industry that feels unlikely. In terms of other input costs, I do not think it is a massive swing factor. You know, at the moment, what is being talked about does not look like it has a big impact on us as a business.
Got it. Got it. Appreciate that. I've been chatting here for about 20 minutes. Want to make sure if there are any questions from the audience so we get those. Microphone right here in front.
At the beginning, you started by talking about there was a period of solving the engineering challenges for the internet and then a harvesting period after we had figured out the right way to solve those. What are the interesting nuances in where we are in solving engineering challenges for corporates right now in AI?
Yeah. Very good question. Let me characterize it this way. November 2022, out comes ChatGPT, the immediate reaction is, "Wow." I, and, you know, I think people felt that it was just going to work out of the box. Very simple implementation. As we went through 2023, people started to realize that there were going to be a lot of challenges around actually putting it into a corporate environment in a safe way. In 2024, a lot was about how to resolve those engineering challenges. Engineering challenges of security. How do you securely put it into your organization, give access to data and so on, without being concerned it was going to be harvested by someone else or ended up being exposed externally in some way.
Regulatory concerns often connected with the similar area, but also some applications where actually you were going to need the regulator to sign off what you were doing. There were then processing cost challenges. The cost of tokens meant that many applications that were viable, so a proof concept would show that the technology works, but actually when you put it through the business case, the cost of paying for the tokens to operate it was more than the cost of actually carrying on with a human-based call center type environment. You got into the engineering of, okay, how do you abstract out the pieces where you really need the AI to be working? You start to minimize your token usage, and do other sorts of processing around it. Even that did not take you to the point.
Now, DeepSeek's come along. You're seeing, you know, that potentially driving a quantum reduction, and that will start to open up some of these business cases that we were struggling with even two or three months ago. The hallucination problem, right? So the reliability of AI to actually, you know, in a customer environment, get it right, get it right every time. We were resolving that through bringing in solutions like agentic AI, where you're essentially creating a team of AI agents who are wearing different hats and they're checking each other, and making sure, you know, one's wearing a regulatory hat, one's wearing a customer service hat, another one's wearing a, you know, produce, produce the answer type hat. Actually, through setting that up, you start to remove the hallucination problems. You layer all these things on.
There've been loads of engineering challenges around implementing AI that we've been working through over the last 18 months or so. Hence, you know, the longer sales cycles, etc., etc. Those are now starting to emerge into business cases that actually work, where the clients have the confidence of starting to move forward.
Let me ask just quickly, we just have a few minutes here. GalaxE, that was a key acquisition for you. You expanded the footprint of talent and skill that you could and your delivery locations pretty dramatically. How's that integration effort progressing, and how is GalaxE growing compared to your initial expectations now?
You want to? It's growing in line, basically.
Okay.
There was initial sort of caution when we bought the business. They're on a, you know, they're integrated in our sort of processes. We've cut over onto our Oracle platform this weekend. They're performing as anticipated. It's very sort of solid. We're very pleased with the progress, actually.
They brought quite a lot of capability to us that fits in.
Yeah.
You know, so underneath that, AI solution that we apply to core modernization, there's a lot of capabilities that GalaxE brought, particularly around being able to read really, really old code and digitize what that code is doing that we can then feed into the AI to get a higher level map out of it.
Got it. Got it. Capital allocation, what color can you give us on timing of share repurchases? What percentage of share repurchases do you intend to finance with cash versus debt? How should we think about that?
We've got a shareholder vote, I think, on the 14th of March.
Okay.
anticipate will go through. It will be a mixture of cash and debt, but one way of looking at it, our leverage, December 2024 was about 0.6%.
Yeah.
I think by the end of June, when I understand we get approved and we start to go through the program, would move up slightly to about 0.7% by June. It will depend on the sort of velocity at which we sort of execute. You know, we may get through it in six months or quicker.
Got it.
We sort of look at the situation again.
What about other use of capital, for example, acquisitions? You know, a lot of times when there's technology change, there can be need or, or a lot, at least a compelling reason to do that. How are you feeling about, about potential acquisitions right now?
Yeah. We would be focused on smaller acquisitions that bring particular technical capability. Some of the IP things that I've been talking about, if added to that, rather than what we've done historically in terms of looking at geographic or sector-based acquisitions. It will be much more focused than it has been.
Got it. Got it. As you wrap up this fiscal year, we kind of are into the headed to the last third of your fiscal year, and you look to fiscal year 2026, what are your kind of key to-dos, from an operational standpoint, John?
The absolute key one is to get these longer sales cycles closing down. That will come with greater assurance around technology and what it can do for our clients and therefore being able to give, you know, those that assurance on the business cases that we're helping them create.
That's great.
If we can bring that down, that then is what leads to the revenue transition into a higher growth scenario.
Great. I appreciate John and Mark for being here today to chat about Endava. Best of luck and look forward to seeing. I think in particular, though, the things that you're talking about where you can develop your IP, increase the speed at which you can deliver solutions to customers and do it in a way that is really beneficial for everybody is really exciting. Thanks for being here today.
Thank you.
Thanks, James.
Thank you.