Curiosity, Trust, and Building AI at Scale — Sudhish Mohan, Group CIO & CTO at TransUnion

0:00If you are not curious, you will be irrelevant because at the speed at which things are moving and the amount of things that are happening around us, including myself, by the way, you would be irrelevant. So you have to be curious because everything you learn now in six months, it's gone. It's gone to something else.

0:21Welcome to Data and AI Mastery, the podcast where we bring you cutting-edge insights, practical advice, and inspiring stories from the leaders shaping the future of data and AI across the globe. I'm your host, Raoul Gabriel-Urma, founder of Cambridge Spark, the leader in transformational data and AI upskilling, career development, and progression. In each episode, I will be diving into real-world case studies of companies harnessing the power of AI to drive innovation, reduce costs, and create new business opportunities. So whether you are an aspiring data scientist, AI engineer, or seasoned executive, this show is designed to give you the tools and knowledge to stay ahead in a world where data is transforming every aspect of business.

1:06Stay ahead, stay inspired, stay masterful. Welcome to Data and AI Mastery.

1:15Hi, Sudesh. How are you? I am very well, Raul. How are you doing today? Fantastic. Looking forward to our conversation.

1:27Great. So, Sudesh, you're the CIO at TransUnion for UK and Europe. I'd love to hear from you. Let's say, you know, you're at a dinner party. How do you describe your role for someone maybe outside of the finance and credit industry? I describe it as one of the most exciting transformation journeys in IT because it brings together a number of things that are very close to my heart. One is technology. I'm passionate about technology. But when you are in the credit industry, you are also making an impact on people's lives, bringing in people who are not financially included in the mainstream of the economy and pulling together those people with technology and driving right outcomes for them is an amazing purpose to have.

2:17So that's what I do. Wow, that's amazing. I can feel the passion. So what does your role involve, you know, let's say on the day-to-day? It would be great to hear it. Yeah, I mean, typically, day-to-day, it is a very operational environment. I remember we are dealing with consumers and banks and businesses across the ecosystem. So there's a large amount of people that interact with our systems. So I'd say that in the majority, it's probably about 30 to 40 % of large transformation initiatives that are happening. But 60 % of the role is making sure that we are interacting with our product organization, with our customers out there to understanding what is happening in their businesses and bringing that to bear on the operations and on our products in our business.

3:15I mean, first of all, being a CEO in such a global organization is really prestigious. So it'd be great to hear what was your journey to this senior leadership position? How did it all start and what was the journey like? It's been a long journey, actually. I actually started out in South Africa many, many years ago. I started working with Oracle at that point in time. And I've had a number of opportunities, but I think two things stand out. One, I've been in multiple industries, from OEM service providers to mining oil and gas companies. I went to Microsoft for a while. I built a digital bank in South Africa, and I came back to TransUnion to financial services.

4:02So it's been a myriad of things on the one side, which gives one perspective across multiple different industries, which is, I think, a fantastic learning opportunity. The second thing is that I've been given opportunity. I've worked with leaders who have given me immense opportunity to grow, to be the leader that I am today. And if it wasn't for those people, I would not be in such a senior role. So I think there's a combination of hard work, there's a combination of luck, being at the right place at the right time, and then just generally doing multiple different things that helps you create a much rounded and shaped way of thinking about things.

4:44I think that, you know, it had to describe my journey in a couple of minutes. I think that would be it. hey fantastic it's great to hear the diversity of sectors i really resonate with that and how that influence your thinking and your career progression and uh surrounding yourself with uh for great leaders you know sounds like you created your own luck so uh that's that's really cool to to see sedish

5:11so let me take you to obviously the the topic of the of the show right let's talk about AI and maybe to kick us off like at a high level how do you think about AI as a CIO as a CIO I know that it's going to be disruptive to our organization I know that introducing it into the organization with the relevant guardrails and that's guardrails around the ethics of what we're doing and embedding some of these new ways of work is a challenge. It's a massive challenge to most organizations out there. I think McKinsey released a report saying that most of these of our AI projects actually are failing or not producing the outputs that were expected.

5:59So should I say it's a slow drip into our organization. And right now we've introduced tools for our developers to start using. A lot of them are using, if I can name a few of them, would be, you know, Claude, the Anthropic model for coding. We have Amazon Code Whisperer at a point in time. We have a number of other things around CoPilot and those sorts of things that the guys are using to do work. Are we seeing the benefits as yet? No. But like I said, it's a slow burn type of opportunity. It'll take some time for people to get used to this new way of work. And I think getting people familiar with the tools, what's happening with the trends in the industry, I think that's an important thing.

6:50And we've taken the first steps into how we start seeing benefits from from ai tools great so you said are we seeing the benefit not quite there yet so how do you think about the benefits in the context of ways of working because clearly those tools especially cogent for software developer has to do with an element of productivity or an element of removing things that are annoying to do like writing tests you know so like how do you of those benefits? So some of the stats that I've seen says there's a 40 % to 50 % developer productivity improvement that people have seen. We have not personally seen that.

7:31We have not, right? But we do believe that it is helping our engineers develop code faster, get to problems faster, and removing the mundane tasks of coding, which is making them more efficient, which ultimately means that we can get them to work on more exciting things and deliver productivity to our products faster which means our customers benefit from that. So that's how we are thinking about it today from a speed of work and the amount of things that we can put out there into the market that will improve significantly. We also don't want our people to be working on lower order activity so to speak.

8:10We want them to be working on the latest and greatest and you know and delivering new features feature engineering new features and getting those things out to our customers and i think it's going to enable us to do that but like i've already stated and i'll make sure i state it again we aren't seeing those benefits as yet it's very early days very early days that's really refreshing uh to to to hear actually so if we think about ai i guess roughly like three buckets the bucket of risk management the second bucket of operational efficiencies and productivity and a third more like customer facing customer delight those three buckets is there like any initiatives currently that you're quite excited about yeah so i'll speak on i can speak to all three buckets so if we look at uh if you look at in the technology world as the cio have mentioned the tools that that we just spoke about that that is the one level the second thing if you look in our data and AI capability, and I run a large part of our analytics business, the opportunity to use models for fraud detection, for example, which is quite big in our business, traditionally models are very, should I say, they're very, very focused on traditional methods of how a system could be broken into.

9:34But using AI and using some of the new graphing capabilities and the new modeling capabilities, you're able to find new vectors of attack that you haven't thought about as an individual, right, or as a human being. And so that is helping us improve some of the models in the environment. If I look at another area of operational efficiency, if you think about our service desk capability where people call in and log calls and have challenges on their credit reports and the like. Using agentic AI in that space to drive out workflows of where things need to flow and logic in there for very, very mundane tasks that we have a lot of people doing, there's certainly going to be some improvement in that area.

10:22So I think in all of these areas, we have some capabilities in play. We need to embed them in the tool. We need to test whether they work effectively. We need to figure out whether these things are really adding value to the environment and we've got to train our users. So those, I'd say, are the vectors that we are attacking in looking at how AI and data is brought into the environment. That makes no sense. So I guess on this specific use case, how do you bridge the gap between I guess the technical problems like improving for detection and the business ROI like is there a way to quantify money saved or customer impact out of that sort of investment so when you think about it from a business perspective we have a current market market share that we own right in in the environment so some of these things we do to ensure that we protect the current market share.

11:20Some of the things we do is to ensure that we are building new capabilities for what is evolving in the market. And that requires some investment up front in order for us to then figure out that at some point in time, the revenue will catch up with the platform that we are delivering. I think what's important in the conversation is that we see there's an evolving landscape out there. We are building the tools and capability, the one true capability, which is a platform that we have in here, that we are trying to figure out that, you know, we are consolidating our data sources on there. We are consolidating our global capabilities to work on that platform and to build these new capabilities.

12:03Those things have a massive upfront cost, right? But over a period of time with what we see evolving in the market, we think is definitely a business case. Is it paying the dividends today? No, it's not, right? We're still building these things out and our revenue is largely from our traditional models, but we see it evolving into a space where we will make revenue from it. And that's how we think about it. We're more into the investment today. I hope you're enjoying today's conversation. If you're finding the insights useful, please do take a moment to subscribe to the Data and AI Mastery podcast and leave us a review on Apple Podcasts, Spotify or YouTube.

12:42Every new follow helps us reach more people and shed incredible work being done by today's Data and AI Leader. All right, let's go back to the episode. I guess I'm curious to hear, like, how do you make the case for budget and investment? Because, you know, there's an element of we can't promise ROI, like in the short term, it's kind of a long evolution of the organization is there any sort of guidance you'd give to you know other senior leaders that might have this sort of similar conversations yeah look i think uh you know in i can't speak to how all organizations work right but there's an element of uh where we create capacity in our business to go and experiment let's experiment with certain technology ideas right?

13:34But what we want to do is make sure we are close enough to our customers to understand the pains that they are feeling. And so if we continue on the path of making sure that we understand what our customers want out there and how the market is evolving, our ability to justify budget to say this is a thing our customer is struggling with. This is what we are building for, right? 20 % of our people's time we can dedicate to building these new things. We know ROI is in the future. And we are testing out some of these things with our customers to make sure that they are seeing value with that. And if we have those necessary data points and metrics associated with what we are doing, the business case in itself is an easy sell because the green shoots are already there with what our businesses are experiencing and our executive sees that.

14:27So I think it's a dynamic environment. We don't know what the answers are, but what we do see is that when we test out ideas like this with the customers, they are buying into it. And so it's certainly something that we use to justify how we get budget for new things. That's super interesting. So there is a saying, you know, more data is always better. And I imagine in industry, you know actually having even more accurate and quality data enables to offer better service to customers so i'd love to hear your perspective around is it true more data is is it always better and then secondly i'd love to hear your thoughts around what makes a good sort of data environment in your world or like what are the things you have to think in terms of infrastructure and so on to make it all work?

15:17So first one, more data. I can't answer. I actually don't know the answer. More quality data assets. Yes, definitely. If you have quality data assets that contribute to the models that you are building 100%, right? In our world, I think there's two worlds here. The one world is regulated data that exists. There's nothing much you can do with that data. It's regulated data that you get, right? that everyone gets. Then there's a second part of alternative data, rental payments, for example, that you could use in the environment to help a lot of people who are not credit visible that you can use to build models with.

15:56And I think if you get data like that, your ability to use non-regulated data to include a large part of the population in with a better model that helps banks and lenders to lend people money because the data that they have gives them a better picture outside of the the control data or the regulated data gives them a better picture that's a fantastic place to be we'd all love love to do that what makes a great data environment i'd say from it was when i think about the data that comes into us and thousands of sources that come in right I want to make sure the quality of the data is amazing. So the onboarding of the data needs to be really quick.

16:45And so right now we're actually using AI on the OneTrue platform to help us get through the data and figure out the veracity of the data and how good it is comparing it to the previous month and if there's any challenges in that data. So that's really a fantastic thing to do. Making sure that data is workflowed effectively into the environment, into the right areas for usage is really important. Some data can't be used in some products, so you've got to be very, very clear about where the data goes. So having the relevant controls in that data is critically important. And the last thing I would say is making sure there's an effective data governance structure of how data is used in the environment, how it is reported, what goes out via product, whether that is effective, you've got to do the checks to make sure that it's working as intended.

17:41And being able to fix any problems in the data when they do occur, that has, should I say, impacted adversely any of our customers out there. So that entire data loop in the environment is really, really important for us. Those are some of the things I'd say you need in the environment. If we take maybe a step back now, I'd just love to hear what areas maybe are keeping you up at night at the moment and, you know, that you're thinking hard about. There's a number of things, right? But the first thing is, if you think about the world that we live in today, every other day, there is someone, there is a cyber attack somewhere, right?

18:24and given the large amount of data that we keep, it's really, really important that our cyber position is always good, right? For lack of a better word, we should be excellent. And making sure that the data that we are custodians of is protected is our number one priority. So that will keep anyone up at night when you think about the size of issues that we deal with and a number of non-state actors and state actors that want want that type of information that will be the first priority the second thing is our operations and making sure that operations are working effectively to cover some of this the data that goes out to our to our customers out there is of good quality and it's helping them make the right decisions because it can adversely impact people and that's what we don't want.

19:20So that's the second thing. The third part I'd say is when you think about our business and the transformation that we are going through, the transformation of our operations, the technology transformation, making sure we are doing all that while we are doing good operations and security is a massive challenge. So we're transforming the business. While all of these things are happening, we also need to keep the business up and running. That's the third aspect. Last one, which I'd call out, is finding talent. Finding talent to make sure that we have the right people who think about these things in the right way and are curious enough to continue building exciting new things, I think is something that's a challenge for many businesses, not just ours.

20:10Well, on the talent side, I'd love to ask you the question because it's quite a unique opportunity to ask that to a CIO, right? So when you look for talent, what are the key traits or attributes that you look for in good talent? Yeah, let me start off by saying that I have a 17-year-old son and I'm trying to encourage him to get into technology and go to a university and do a degree and all these sorts of things. But really, in my heart of hearts, I am not sure whether I want him to do a degree because in three years, it will probably be obsolete, right? And the practical experience of being in a business, I think, is more important these days.

20:54Now, I wouldn't tell him that because then it would be an excuse for him not to want to do a degree. And I say that because when I recruit people, I think, no, I have a degree in this. I don't think those things are really important anymore. What I look for in people is curiosity, work ethic, curiosity, multiple different interests in how they think about a problem. I think that's really important. The next thing which I think is important is having people skill, the ability to collaborate, have a conversation, understand each other and find solutions for the business. I think I look for people like that.

21:38Now, you can't really judge a lot of that stuff in an interview, right? It's quite hard. But, you know, when people can demonstrate that they are collaborative and the ability to find solutions, find a halfway ground, so to speak, I think are really important for me. That's what I look for. I don't look for qualifications if that's what you we're going towards because I don't think those things are relevant anymore. I like what you said that, you know, curiosity is an important attribute, especially with AI where you get a bunch of stuff that may be right or may be wrong. So kind of questioning it and being curious.

22:14It's really important, but I think Steve Jobs said it in his Stanford address some many, many years ago, be curious, you know, whatever. And people just take that for just for the word that it is. But I was saying this to my team the other day, actually, we were speaking about AI. And I was saying that if you are not curious, if you are not curious, you will be irrelevant. Because at the speed at which things are moving and the amount of things that are happening around us, including myself, by the way, you would be irrelevant. So you have to be curious because everything you learn now in six months, it's gone.

22:50It's gone to something else, right? How are you going to keep ahead of the pack if you need to go to university every time? No, it's got to figure out better ways of learning and being ahead of the pack.

23:06Hey, I'd love to take you to maybe a quick fire round of questions. I guess one, you know, you have a unique perspective as a CIO. So is there any contrarian view you have in the industry, something that's acknowledged, but you kind of would challenge it? I think cloud should be a tool in the capabilities that you have to deploy work. It shouldn't be the only tool that you should deploy work. I think as we are starting to roll out more into cloud, and when you're not a cloud-first business, you'll find that the costs become quite exorbitant over a period of time, and it starts to hit quite hard.

23:50And I think there are companies that are rethinking about rethinking their models about moving to cloud. So I don't believe everything should move to the cloud. I think we should move certain things to the cloud for maybe analytics capability or whatever the capability is, right? But I think we should be very, very cognizant of making sure that the costs don't outweigh the benefits that we get from it. and we should, in my view, in my view, certainly in my industry, I'd rather be pushing for a hybrid solution where quite a bit stays on-prem versus in the cloud. That's the only one thing I'd say my mind is shifting over the years.

24:35That's a great contrarian view, which is really refreshing and you're absolutely right. There's a big push. Everything needs to be in the cloud, but actually the cost argument is a great one. it can be quite expensive to just put everything out there so bring on-prem the next question i have is you know clearly it's moving so fast so i'd be curious to to hear for the audience like how do you stay up to date as a leader like what are your you know like networks or platforms or sources to kind of like you know stay up to date in this fast moving world i'm very lucky that i go to leads every week uh by train which is about two hours so in those two hours i make sure that i've downloaded enough podcasts to keep me busy for the two hours there and two hours back.

25:20I'm making sure that I'm reading the latest trends of what's happening. I am trying to educate myself much more in two areas. So stuff that I said is quite general, but the other two I'd say is making sure that cybersecurity at a board level and how I'm representing the company at a board level is really important. So I keep up to date on readings on that. I'm a member of the Institute or directors in London. And there's quite a few things that they send out that I keep ahead of. And then I am also quite a big fan of Coursera and of deep learning AI, the Andrew Ng website, should I say. And whenever they come out with new stuff, if it's within my technical remit, because some of the stuff gets really complex, even like I have no understanding what he's talking about.

26:09I do try to keep ahead of stuff over there. So that's what I tend to do. I do read a lot. I'm an avid reader, lots of articles, lots of books, books more in quantum science, things that I'm interested in. But that's what I tend to do a lot of. Great, great, great advice. So I've got two more quick questions, a bit more personal. What was your favorite subject back at school? Oh God, this is a terrible question. Because every time I tell my children history, they go like dad why do you want to know about the past right uh my favorite subject in school has always been history actually my pastime is reading history particularly roman greek all of the different empires that rule the earth i'm fascinated by that stuff i read a lot about it my favorite and final question i guess picturing you on this two hours journey to to leads what's your favorite music genre to listen to?

27:11I don't listen to music. Just podcast. I just listen to podcasts. I do not, to be clear, so I listen to only things that are factual. I only read non-fiction books. I do not listen to music, although I would tell you I went to the Beyonce concert in London with my daughters. I was bored out of my mind. I sat reading my books while I was sitting there. And next year, I've already bought tickets for my daughters to go to the weekend concert in London. And they've begged me, they've begged me to please keep my phone at home because they want me to enjoy the concert. So unfortunately, I'm quite boring like that.

27:56I don't listen to music. Hey, well, I look forward to hearing how the weekend concert goes. I mean, there's some good tunes there, so I can't wait to find out. So Sudesh, it's been an absolute pleasure to welcome you on the show. Thank you. Thank you. It was very good. Challenge you, Raul. Appreciate the time.

28:17Super conversation with Sudesh. My favorite moment was when we discussed about, you know, what's keeping you up at night as a CIO? Like what are the things that you're having to think about? And there were like four things, you know, obviously cybersecurity, big topic, especially now with AI. because you can really scale attacks and make them more intelligent. So that's kind of like a bucket that you're having to think really hard about. The second one was around operations, data infrastructure, data quality, ensuring that it's as good as it can get. The third one is really the transformational agenda.