How Endava is Helping Companies Turn AI Investment Into Measurable Business Value

[00:00:04] - [Speaker 0]
What does it actually mean to become an AI first business? Spend five minutes at any tech conference and you'll hear vendors promising that AI can transform productivity, cut costs, automate workflows, or even solve just about every business challenge imaginable. But once the keynote ends and the sales pitches stop, many businesses are left wondering why the results aren't matching some of those big promises. Well, my guest today has spent the last few years helping organizations answer that very question. His name's Matt Cloak.

[00:00:42] - [Speaker 0]
He's the CTO at Endava where he's helped steer an 11,000 plus person technology business toward becoming AI native. So rather than treating AI as just another software rollout or chasing the latest shiny tool, Matt believes the biggest opportunities come when AI becomes almost invisible, woven naturally into the way that people work every day. So today, we're gonna talk about why AI hype is colliding with enterprise reality, why buying thousands of licenses seldom delivers the outcomes companies expect. And why the organization seeing genuine success are the ones that are focusing on mindset, culture and workflow redesign rather than just technology alone. And if you've ever wondered why some companies seem to be getting real value from AI while others are still stuck in pilot mode, I'm hoping that today's conversation will offer you practical lessons that you can take back to your own organization.

[00:01:46] - [Speaker 0]
But enough from me. Let me introduce you to Matt right now. So thank you for joining me on the podcast today. Can you tell everyone listening a little about who you are and what you do?

[00:01:59] - [Speaker 1]
Certainly. So my name is Mac Cloak. I'm Endava's CTO. I wear a couple of hats within the organization. So I'm responsible for our internal technology, team, which means like every other CIO out there, I'm kind of like frustrated by the general state of our IT and are we moving and changing things quickly.

[00:02:17] - [Speaker 1]
I'm responsible for defining our capabilities. So what are the things that clients want to buy from us from a technical perspective and the skills required to actually go and make us a successful IT services company. And then last but not least, I'm the opinionated individual when it comes to, well, how can we embrace technologies like AI and, really make a difference both for our clients but also for ourselves. And I'm sure we're gonna talk about it, but, I always consider ourselves client zero when it comes to AI adoption and AI mindset.

[00:02:51] - [Speaker 0]
I love it. And we're recording this when I finally get a few weeks break. I've been to thirteen, fourteen different tech conferences already this year. And predictably, everywhere I look from Egypt to Vegas, it's vendors. Or every vendor is promising AI or agentic AI as the answer to just about every business problem.

[00:03:09] - [Speaker 0]
I've got to ask, from your perspective, where is the gap between those big promises being made on the show floors and keynotes and the reality the organizations are experiencing right now?

[00:03:20] - [Speaker 1]
Yeah. I mean, AI fatigue is real. I can see the headline now across many blog posts. The way I look at this, and I said earlier about being client zero, I think as an organization about two, two and a half years ago, we had a bit of an existential moment because everyone started looking at ChatGPT. I mean, it was GPT-three at that particular moment in time and was kind of, well, I don't think we need IT services companies anymore because we're just gonna get the AI to do everything that you were currently doing.

[00:03:53] - [Speaker 1]
And the first two or three times you heard that, there was a kind of like, I don't know what you're talking about. And then you carry on a little bit further and you're like, actually I think there is a bit of a thing that we have to address with this. So to that whole thing of everybody and their father promising that AI is going to solve all problems, For me the most important thing is talking to people who have credibly gone on that journey. What have they done for themselves? What is the thing that they've done?

[00:04:21] - [Speaker 1]
Where they can tell you what is the actual experience they've got. There are plenty of people out there floating around with PowerPoint decks and Keynote slides all saying, this is the wonderful thing and this is gonna happen for you. But turn around and go, so so what agents are you as an individual running? How is your company actually going on that journey? What have you changed as a result of AI, coming into the world?

[00:04:45] - [Speaker 1]
And you very, very quickly, sort out those who are would be AI native people versus people, like ourselves who've been on that journey for a couple of years.

[00:04:56] - [Speaker 0]
And before you join me today, was reading that you've spoken often about becoming an AI native rather than simply adopting AI tools. So what does an AI native organization actually look like in practice? And how is that different from just layering AI on top of some existing processes that you may have seen?

[00:05:14] - [Speaker 1]
So so I love telling stories. So so I'll tell you the story about twelve months ago. I was at an event up in the French Alps. It was a beautiful it was like a James Bond villain retreat. Was that kind of thing.

[00:05:28] - [Speaker 1]
And and there were many CEOs at this particular event that I was talking to. And I was talking about the journey of becoming AI native, what did it mean to me. My take on it is it's about curiosity and thinking about how can AI help me solve a problem as the first thing I think about as opposed to it being something that's bolted on. So I gave this talk and I was talking to the CEO and the CEO turned around and went, well, I've done my AI transformation. I've deployed 10,000 licenses of ToolX to my organization.

[00:06:00] - [Speaker 1]
You should come to me and I'll tell you how it is to become an AI native organization. Anyway, roll the clock forward twelve months and that CEO is no longer with that firm. And that firm has now come back to us and said, could you come and help us explain how to become AI native? Because we deployed all of these tools and we saw little to no return on the investment that was made into it. So for me, it's a my AI native is a mindset.

[00:06:26] - [Speaker 1]
It's really about that kind of like curiosity. What can I do? How can I do, how can I think about using AI on a daily basis to help me?

[00:06:35] - [Speaker 0]
What a great story. Absolutely love that. And another one of your more most interesting views that I was reading about is that, AI is best when it's invisible. So can you tell me a little bit more about what that means and why successful adoption might actually involve people noticing it less, not more, and waving the flags from from their offices.

[00:06:54] - [Speaker 1]
Exactly. So I think I started talking about invisible AI when and again, it goes back a couple of years when people were kind of like, I I wanna be able to see and touch AI before I know that AI is actually happening. And it was kind of like, well, you're already using AI because if you take a picture on your phone, that's literally using AI and other processing before you ever actually get to see the image on the screen. And, you know, oh, I wanna remove that tree branch coming out the side of your head, and I'll just, like, scrub around it and it will disappear. That that that was the beginning of this idea of AI being invisible, as in it's just a thing that happens without you going, that that's a really clever thing.

[00:07:34] - [Speaker 1]
And again, I think that people have been on a journey with AI, and the chat interface has been great. It's democratized the idea that someone can type away to a little assistant and say, could you help me with this? Or could you review or summarize this document? But where we're moving to now is agents being able to surface things for you without you ever really having to think or direct them to do it. So to give you an example, I was away with the rest of the executive team last week.

[00:08:04] - [Speaker 1]
And I knew that I wouldn't really be able to answer my emails or my team's messages for a couple of days because I was going to be very focused on a particular activity. So I just fired up some software agents to keep an eye on my inbox. And it was like, if you're getting this type of message, then redirect it to these people over here. If something urgent comes through from this group of people, then break through and kind of alert me. So what happened was every time that I came out of a break, all I had to do was go into my Drost folder and go send, no, delete, that's not oh, that's not quite my wording, and I can do it.

[00:08:37] - [Speaker 1]
And that's an invisible use of AI. I, you know, I wasn't sitting there typing away at the keyboard waiting for it to happen. I just had the agents working on my behalf in the background.

[00:08:47] - [Speaker 0]
Wow. That's incredibly cool. And everything you just mentioned there will resonate with so many people thinking, hey. I could use something like that. But at the same time, they're scared of handing over their their inbox to an agent.

[00:08:58] - [Speaker 0]
Does it did you go on a journey yourself there? Or or it sounds like you added it to draft first. You just let it blindly send stuff.

[00:09:05] - [Speaker 1]
Yeah. I mean, we we talk a lot again to that AI mindset, that AI nativeness is about the responsibility of the human in the loop. Now human in the loop doesn't necessarily mean that you're micromanaging every decision that's being taken by AI, but it is about understanding when is the most appropriate time for you to kind of like step in and intervene. I'm not quite comfortable yet that it's going to send an email directly from me, but drafting a response, it's a lot quicker for me to go, yeah, that sounds what I wanted. Over a period of time, it's got to know my tone of voice and the use of smiley faces and other things in the way when I send out outbound emails.

[00:09:47] - [Speaker 1]
But again, I think it's you know, the the way I think about this is if you're a leader or a manager of a large team, you don't micromanage the outputs of absolutely everything that that large team does. But you think about, well, how do I supervise this group of people so that they're doing the job that I've asked them to do and they're not doing something stupid? And I think that's fundamentally how you should think about agents is I'm not gonna review the output necessarily of every single decision it makes, but actually, I'm I'd I'd there are some things I will step in and go, no. I wanna see the output of that email before I send it.

[00:10:23] - [Speaker 0]
And I think you're in somewhat of a unique position here and a unique vantage point because at Endava, you work across multiple industries and indeed markets. So I've got to ask, what are some of the common mistakes that you see organizations making when they approach AI as a standalone initiative rather than a broader business strategy, and maybe they're just trying to tick that AI box or be part of that AI narrative? Do you see a lot of similar mistakes?

[00:10:48] - [Speaker 1]
So so I see two, and some of them would definitely get people to be like angry man shakes fist at screen side of things. The the the number one problem is seeing AI tooling as a rollout problem. So I have put all of my developers, GitHub Copilot or Cursor or Windsurf or Devon, whatever is the tool du jour, why am I not seeing returns? And again, it's kind of like, well, there's a guy called Jeremy Upli from the Stanford D School, writes some brilliant newsletters around AI. And he talks about that's defensive mindset that's concentrating on capturing 10% to 15% improvements.

[00:11:31] - [Speaker 1]
It's not about value creation and reimagining a process using AI as the basis of doing it. So tool adoption is definitely one where people are like, if you're just buying tools and handing them out, then please don't expect significant changes in what you're getting back in return. The other one that makes me shudder and the bit where angry man will shake fist at screen when I say this, AI centers of excellence. This idea that there will be a group off to the side that somehow will sit there, stroke their beards like myself and kind of write things down and then pass them off as sage advice into organisations. Everybody across the organisation needs to be challenged in terms of how are you going to use AI in your context.

[00:12:20] - [Speaker 1]
How can you have a champion within your team that's going to move this agenda forward? Our legal team's been brilliant at doing that. I mean, they now have a monthly hackathon where they are self organizing around the concept of they have to improve their processes using AI. As a technology organization, we help them at the beginning by introducing them to the right training and everything else. But, you know, I wouldn't pretend how to tell a lawyer what is the best way of using AI.

[00:12:49] - [Speaker 1]
So, yes, AI center of excellence is make me shudder every time I see a PowerPoint slide with one.

[00:12:56] - [Speaker 0]
Yeah. I can hear so many people nodding their head in agreement there listening to you. For people listening that are in highly regulated sectors such as payments and financial services, reliability and trust, these things are nonnegotiable in these industries. So what lessons from running mission critical systems that you've learned? Should should business leaders apply to their AI strategies?

[00:13:17] - [Speaker 0]
Any advice or anything you'd pass on there?

[00:13:20] - [Speaker 1]
Yeah. I mean, we've spent a lot of time over the last, twelve months codifying how to use AI in what we refer to as an engagement life cycle. So in that very first conversation that you have with a client all the way through to helping them understand their business problems, exploring the solution space and doing it, actually building solutions for people. And we've codified that in something that we refer to as Dava Flow. And within Dava Flow, we are able to make it work in a regulated environment because what we do is we go, give us your requirements.

[00:13:54] - [Speaker 1]
If you are working in a regulated environment and you've got an auditability requirement or your traceability requirement or maybe it's a recovery objective, then let's codify those things as policies that the AI has to adhere to and demonstrate that it's meeting those policies as you're actually going through the process of using them. So what we've learnt is that effectively what are non deterministic systems, but guess what, humans are non deterministic as well. How can you actually use that and put controls and policy around it so that you can get the outputs you want while still adhering to all of the requirements from a regulated environment?

[00:14:33] - [Speaker 0]
Do you need AI agents that you can trust? Well, with an AI data layer providing real time connection within your data platforms, you can trust your agents to provide accurate solutions. So scale your business by trusting your agentic AI, accurately getting the work done for you. Trust its capabilities with Denodo. And you can do that by simply visiting denodo.com to learn more.

[00:15:01] - [Speaker 0]
And I think many organizations are all searching for that breakthrough AI application that will transform their business overnight. And just saying that out loud, maybe they're looking in all the wrong places for that. But, mean, are companies focusing on the wrong goal, and and should they be paying more attention to incremental improvements across the organization instead? Again, any big mistakes that you see here when they say, I wanna do AI?

[00:15:25] - [Speaker 1]
So, I I think I've actually, written this down in a few posts which are out on the Internet if you search my name, which is, don't believe in the silver bullet. I mean, I think the many, many organizations are enamored of the idea that if there's just this I could just implement this one system with AI, then that's going to make all of the difference to everything that I do. So everyone was looking for the killer application. And it's like, well, if you're talking about productivity and efficiency, you can see absolutely massive changes by thinking about how can I use AI in an end to end workflow without ever actually having to write a new system? I mean, to my example earlier, and it's a trivial one about the look at my inbox, draft my responses to me, just generally keep an eye on the shop whilst I'm stuck in a workshop for two days.

[00:16:17] - [Speaker 1]
Well, think about that example from our legal team in terms of well, if I've got an agent looking at an inbox that's gonna read an NDA, which is gonna use some form of, knowledge or prompt to validate that the NDA, falls within bounds, adheres to x, y and z terms and is able to draft a response, then I freed up that lawyer from having to do that deep dive into there so that they can concentrate on higher value work. I've not written a new system. I've literally just given them a way to use the existing tools in a more efficient way. So so my philosophy around not looking for a silver bullet is if you're if you're thinking that you're, you know, gonna pick up this alien artifact, drop it into the center of your organization and go, duh duh. There you go.

[00:17:02] - [Speaker 1]
Success. Let's all go home. It's just not gonna happen.

[00:17:06] - [Speaker 0]
So if we zoom out for a moment, how do or else should a leader decide where AI belongs within their business? Are there any particular workflows, processes, or decision points where you've constantly seen the strongest results? Because I think there's a big focus on ROI, business outcomes now. Are you seeing different areas perform better than others?

[00:17:27] - [Speaker 1]
I think I'm quite old. For those who are listening in there as opposed to seeing the video, I feel very long in the tooth now. I've been doing this job for over thirty years in one way or another. We used to talk about value chain analysis, which was this idea that you would look at how something is done from an end to end perspective and you'd look at, you know, information is being handed off between which parts of a process, what systems are involved, etcetera, etcetera. And when you were able to draw these pictures of like value chains of saying how do we do x inside organizations, they'd invariably create these light bulb moments for people to go, well, I didn't realize it was that complicated or I didn't realize I had that dependency on that particular system.

[00:18:11] - [Speaker 1]
I think what AI does is allow you to go through from that kind of like visual comprehension of what's there to begin to then focus in on going, actually, where can I make this process more efficient? If literally I'm sending something off to one system to be validated by a human and then something comes back in a slightly modified form, is that an opportunity for us to look at codifying that, putting in an agent, doing something that allows me to make that overall process? So I wouldn't say that it's like a particular system where you go, oh, inside banking, it's the core banking. I can make that 10% more efficient. I think it's looking at all of those workflows that humans have created around those systems and doing what a a good old fashioned value chain analysis would do and go, why do I send that email to Bob?

[00:19:03] - [Speaker 1]
And why does Bob then send it to Sharon? And then when Sharon sends it back to me, how different is it from the original thing I sent to Bob? So that's a lot more about how you can start thinking about those incremental improvements.

[00:19:15] - [Speaker 0]
And as AI continues to be embedded into our everyday business operations, what role do you see human oversight, that human in the loop playing? And and how do organizations strike that right balance between automation but also keeping accountability there as well?

[00:19:31] - [Speaker 1]
Yeah. I mean, it it it it's probably the subject of our lifetime and the subject of, on everyone's lips, which is what happens to me and what happens to my job if we start deploying this technology more broadly. I tend to be an optimist, which I think is work changes, it never actually goes away. So therefore, even if we look at that process of Bob and Sharon and everything else, for the individuals of Bob and Sharon, who knows what higher value work they can do inside the organisation if I take those two relatively menial jobs away from them and give them something else to do. And one of the other things that I think, and I use an analogy, so again forgive me for my terrible analogy, when all of the New York high rise buildings were first built, there used to be a human operator of the lift.

[00:20:25] - [Speaker 1]
And you'd go in there and you'd go, I'd like to go up to the twelfth please. And they'd kind of like press all the buttons on behalf of the people and turn a handle and they'd go up to the 12th Floor. And over a period of time, those were systems that were effectively handed over to machines and lift control systems. And now we have those infuriating lifts where before you even get inside them you have to say, Where are you going? And they tell you which lift you should be going up.

[00:20:48] - [Speaker 1]
But that's a rant about lifts. But the point being, for the people who operated the lifts, that job didn't exist anymore. But other jobs were created by building all of those high rises and building all of that infrastructure to be able to do things. And then you turn the circle all the way to the end, and you're kind of like, and now you do see humans back in the lifts when you're looking at very high touch, high value, high service apartments and office spaces. You see people back in those roles because the connection to a human becomes important again.

[00:21:21] - [Speaker 1]
So it's a terrible analogy about lifts, but I think it's more work shifts as opposed to disappears.

[00:21:28] - [Speaker 0]
No. I think it's a great analogy. And if we look back, I think we've seen so many changes in our work lives in the last three to five years. If the pace of that seems fast, there's a danger that it'll never move that slow again. So if we look ahead to another two or three years ahead, what do you think will separate those organizations that successfully integrate AI into their culture and operations from those that just spend heavily on AI but struggle to show any meaningful business outcomes like that person you referred to earlier?

[00:21:58] - [Speaker 1]
Yeah. I I I have a perspective that basically says the companies that will be successful in the future are the ones where the leaders demonstrate how they use AI technology and how they then take that into their organisations and make them successful. So it's my anti pattern of the AIC at COE. That's like a senior manager waving their hands going, oh, I've got absolutely no idea how to do this. So I'm just going to create a committee and they're gonna go and solve the problem.

[00:22:28] - [Speaker 1]
As opposed to a senior leader going, actually I got curious and I used AI in this particular way and I saw the value of it, therefore we are going to go and be successful because I'm gonna model the change in the behavior that I want to see within my organization. A slight tangent, couple of weeks ago I took our chief people and location officer, the chief of staff to the CEO, the head of general counsel and the CIO who works for me and a group of others. And we spent a day writing software agents and being shown the art of the possible with one of our partners, OpenAI. And that was an amazingly powerful experience because people were like, oh, I can actually write this thing for myself. And, if I do x, y, and zed, I'm kind of like, oh, there were lots there were so many light bulb moments during that session to the extent that the CIO turned around to me halfway through the morning.

[00:23:22] - [Speaker 1]
I was just like, you know that contract renewal we were talking about, VendorX. I don't need Vendor X anymore because I've just worked out how in this room I can do the thing that I was going to pay Vendor X to do. And again, model the behaviour at the top and then push that down inside of an organisation, you'll be successful. Don't defer the responsibility to people somewhere else to come and tell you how the world is gonna change as a result.

[00:23:47] - [Speaker 0]
Wow. I think that is a thought provoking moment to end on that. I know you do write a lot on LinkedIn, etcetera. So for anyone listening, wanna keep up to speed with some of your thoughts and musings and the kind of things that you're producing or find out more information about Endava and all the ways that you're helping your customers and clients, etcetera. Where do you like me to point everyone?

[00:24:07] - [Speaker 1]
I I would say LinkedIn. I can't say that you won't get some updates with spelling mistakes and grammar mistakes, but that I should probably get AI to review my outputs a little bit before I put them up there. Also, check out our company website as well, so endava.com. I believe we're gonna be doing some fun and exciting things on that over the next couple of weeks and months as well.

[00:24:29] - [Speaker 0]
But I think anyone that goes on LinkedIn now, and you see all those statuses that all look the same with three, four word sentences, soon have a few spelling mistakes in the generic AI posts. Would you agree?

[00:24:41] - [Speaker 1]
I I did. LinkedIn is something that, I I think I probably signed up to LinkedIn within the first couple of months of it coming into being. I kinda it's so far in the distance. I can't actually remember when I did it. And for a very, very long period of time, I wouldn't accept any connection unless I actually had physically worked with the individual.

[00:25:01] - [Speaker 1]
So it was that thing of kind of like, you know, it was a very, very curtailed list of things. Nowadays, it's a full blown social media network, isn't it? And and if you're not keeping up with it, if you're not feeding it, if you're not doing things, then then somehow you feel that, you're not fulfilling your obligation as someone on LinkedIn.

[00:25:21] - [Speaker 0]
Love it. Well, I will add links to your LinkedIn, Endava's LinkedIn, and the website and everything else there. And I just love chatting with you today about why AI hype is colliding with enterprise reality, AI works best when it's invisible, and how it's not about those killer apps. It's actually about industry wide integration or company wide integration. So I urge everyone listening to check out those links.

[00:25:44] - [Speaker 0]
Keep up to speed with you. But more than anything, just thank you for bringing this conversation to life today. Really appreciate your time.

[00:25:50] - [Speaker 1]
No. It's been great, Neil. And look, if you wanna have a chat in the future, you know exactly where I am.

[00:25:55] - [Speaker 0]
Wow. I think there was so much to take away from my conversation with my guests there. But I think one of the things that kept resurfacing throughout the conversation was successful AI adoption has far less to do with finding a magical tool, a silver bullet, and far more to do with changing how people think about work. And I think Matt shared some fascinating examples there of AI operating quietly in the background, helping people make better use of their time without demanding constant attention. And he also highlighted something many organizations are learning the hard way, and that is rolling out AI tools.

[00:26:34] - [Speaker 0]
That's the easy part. But changing behaviors and processes, that's where the real work begins. And maybe the most interesting insight of all was his view that leaders cannot simply delegate AI transformation to a committee or a third party vendor and just hope for the best. The organizations that thrive are likely to be the ones where leaders are actively experimenting, learning, and demonstrating what's possible for themselves. But I'd love to hear your thoughts on anything we covered today.

[00:27:05] - [Speaker 0]
Are you seeing AI become a natural part of everyday work inside your organization, or are you still trying to bridge the gap between enthusiasm and measurable business outcomes? As always, you'll find links to Matt, Endava, and the topics we discussed in the show notes. But if you got anything to talk with me about, you wanna work with me, just have a little chat with me. Techtalksnetwork.com. 4,000 interviews over there as well, and you can leave me an audio message.

[00:27:35] - [Speaker 0]
But that is it for today. So thank you for listening. I'll be back again tomorrow with another conversation like this one, and hopefully, you'll join me again then. Bye for now.