Episode notes What happens when a legacy institution in a heavily regulated industry turns AI from a compliance challenge to a competitive advantage? Our guest for this episode, NatWest Bank's Diana Kennedy, shares how the financial services leader does exactly that. The wide-ranging chat explores how NatWest Bank approaches AI-driven transformation, its approach to partnership, and expectations for leaders in setting the pace for innovation. Featured experts Diana Kennedy, Director, Architecture and Engineering, NatWest Digital X Ralitsa Nenkova, Global Insurance Leader, Consult Partner and Vice President, Kyndryl More to the story Scaling AI in regulated spaces Read Diana Kennedy’s featured article from the Kyndryl Institute on finding the right partners for AI at scale. AI in Insurance Explore Kyndryl’s Insurance Recommendation paper for applied industry insights.
Episode notes
What happens when a legacy institution in a heavily regulated industry turns AI from a compliance challenge to a competitive advantage?
Our guest for this episode, NatWest Bank's Diana Kennedy, shares how the financial services leader does exactly that. The wide-ranging chat explores how NatWest Bank approaches AI-driven transformation, its approach to partnership, and expectations for leaders in setting the pace for innovation.
Featured experts
Diana Kennedy, Director, Architecture and Engineering, NatWest Digital X
Ralitsa Nenkova, Global Insurance Leader, Consult Partner and Vice President, Kyndryl
More to the story
Scaling AI in regulated spaces
Read Diana Kennedy’s featured article from the Kyndryl Institute on finding the right partners for AI at scale.
AI in Insurance
Explore Kyndryl’s Insurance Recommendation paper for applied industry insights.
Tom Rourke 00:03
Welcome back to The Progress Report. I'm your host, Tom Rourke. Financial services are critical to the global economy—often an early adopter of technology—but it’s also an industry that’s governed by tight regulation and, in some instances, a certain level of risk aversion. However, I’m delighted to be joined today by two very experienced practitioners from the industry who are going to tell us why this is the most exciting time of their careers: Diana Kennedy, who is the Director of Architecture and Engineering at NatWest, and my good friend and colleague, Ralitsa Nenkova, who is our Global Insurance Lead. So, Diana, Ralitsa, you’re both very welcome to The Progress Report.
Diana Kennedy 00:38
Oh, thank you. Thanks for having me, it is so very great to be here.
Tom Rourke 00:41
I've got to jump straight in, Diana, maybe just to ask you a little bit about your own journey into this world of agentic AI, but built on a whole load of foundations in architecture and IT.
Diana Kennedy 00:52
I really genuinely think it's the best time ever to be in technology, even through my whole career. So I've been a technologist, with a whole career engineering degree, but worked in enterprise technology, a bit of time in consulting, actually in the financial services sector. But in the last long while leading the design, build and implementation of enterprise scale technology, and that's been in the oil and gas sector, that's been in healthcare, and now in financial services and I really think with our world of AI and now agentic AI, the foundations of enterprise technology are even more important than ever, getting our data platforms right, making those that data accessible via API's building and scaling the end user capabilities for our colleagues and customers to access that that data and use those models, and now critically thinking about the skills that we need as an organization and at every level, and how we then make AI accessible to our customers in a way that's really meaningful and helpful to them.
Tom Rourke 02:06
I’m sold, but I was actually smiling because I think your introduction is a fantastic reminder of the importance of fundamental architecture and that view, irrespective of the pace and novelty of change. I would argue that there are still some people who need to be persuaded, but maybe we can help with this discussion. So, Ralitsa, you and I have a shared background in insurance. Is this the most exciting time to be in insurance and insurance technology?
Ralitsa Nenkova 02:32
Absolutely, yes. And I've never been more inspired actually by how insurance has responded to it than overall financial services. Because Financial Services is built on hindsight, cautiousness, and risk management. There were conversations 10 years ago where you couldn't persuade an actuary to do anything unless they had 10 years of backdated data. So we’re moving to proactivity, foresight, immediate insight, and decision-making by agents. That's a massive leap. And for those that are away when insurance was born, which was when Lloyds of London was set, that is very many centuries ago, we haven't changed our operating model since then. And now is the time, and it's really happening, and everybody is doing it. So yes, it is the right time, and this industry has never needed it more.
Tom Rourke 03:23
I'm smiling because obviously when people talk about financial services and they take banking and insurance together, I always smile. Well, banking has this facility with risk, whereas insurance is about avoiding it. So here what you're saying. It really is quite a change. And so, so let me just talk move a little bit on to the kind of practicalities. And it feels that, you know, there's still a huge amount of conversations, moments around people with a real anxiety around, are they realizing the benefits from their investments, right? And later, we'll talk about how you measure that. But maybe if you could just talk a little bit about NatWest's experience, because it feels as if you've managed to put early scores on the drawer.
Diana Kennedy 03:57
We see AI across three pillars. Firstly, AI for colleagues, and we take that responsibility really seriously. Actually, as a UK bank that is very embedded in the UK economy, we process a significant proportion of the UK’s payments every day. We bank for the UK Government. We bank the royal family. Actually, we’re very embedded in the success of the UK economy. Therefore, the social responsibility we have to upskill our colleagues—60,000 to 70,000 of them—in the use of AI and remove nervousness and uncertainty is one that we take really seriously. So we’ve rolled out AI agents and Microsoft Copilot to every single employee, including all of our frontline workers, including all of our back-office agents in India and other parts of the world, really, as a means of encouraging them to experiment; to think about their own personal productivity; and to create small automated workflows with some intelligence that wasn’t previously possible. We see really extraordinary adoption, actually. And I think if you create an environment from the top, from CEO downwards, that is really clear but visionary in terms of the use of AI, that encourages everybody to participate. The second is AI for customers. We now have a large number of projects—about 600 AI projects in the bank—over half of which are specifically targeted at our customers and building and improving the customer experience: building new products that are anchored in AI. That shows up through the people that we’ve hired. So we’ve hired the AI research team from Meta; that whole team is now in NatWest, and they’re investing and researching in fraud detection, deepfake detection, and a number of other things. And then the third pillar is AI for process. So we, of course, see the opportunity for AI to simplify, modernize, and reinvent how we do work. And I think that’s where NatWest particularly is leading the industry: thinking about AI as a mechanism to reinvent—not just to compete with the more cloud-native neobanks—but actually to exceed the expectations of our customers and new customers too, and how we can really make those experiences fantastic for your financial wellbeing—things that are traditionally quite difficult to do or awkward, or you have to phone up or go into a branch, or whatever. I think that’s the joy of AI right now: you can have simplification and removal of friction, and you can make it awesome.
Tom Rourke 06:50
Yeah, I particularly like that, not just automated, but fantastic. But it is interesting. If you think about financial services, you know, you describe phenomena infectious enthusiasm around all of those things for your customer and your colleagues. But obviously there's also a context where you don't have complete liberty in terms of the reinvention processes, because there's a regulatory context, and maybe relative because obviously that spans also insurance that that you know, there are both opportunities which we might get to but also, I suspect, some moderating factors when it comes to regulation in a regulated industry on AI. So maybe we could talk about that a little.
Ralitsa Nenkova 07:24
Actually, regulators across the globe—not just the UK—are very encouraging of AI adoption right now, because this is the only way we’re going to drive customer experience outcomes that we’ve never seen before and maintain growth in the industry, given the other challenges in the wider economy. I think the main part around “yes, it’s a regulated industry” is what guardrails you put in, how you adopt, and thinking through—at the design stage—all the controls and guardrails required to prevent challenges further down the line. So it’s a controlled and safe adoption of AI versus experimenting without any particular vision. When you look at different industries, particularly insurance, awareness is very high; consensus is there; adoption is still shallow and a little bit slow. So moving from this proof-of-concept (POC) position to adoption at scale and production—which Diana was giving fantastic examples of how NatWest has done—is still more the exception than the norm, and I think everybody’s learning from each other. But we’re expecting significant acceleration. Actually, this year is the year of agentic AI at enterprise workflow level, not at POC level. The other barrier at the moment is the operating model and people. And maybe we can dive a little bit more into people, because if those two things are unblocked, then really it’s unstoppable—there are no limits to what we can achieve and what difference it will make, actually, to the financial wellbeing of everybody around the world, which is significantly required given the geopolitical challenges. So that would be my view: actually, regulation is a positive, and it gives us the impetus to do it in the right way.
Tom Rourke 09:13
You touched on the people thing, and I was struck by a phrase you used in an article you wrote with the Kyndryl Institute recently, around the idea of hiring people who are “wildly competent”. It strikes me that one of the things we need to start thinking about is talent acquisition: having teams that actually know how to look for that. It’s more than just writing a specification—there’s a deeper conversation needed. Are you seeing people from very different disciplines coming into NatWest? Are you also having to think about how you identify those people within your own organization?
Diana Kennedy 09:49
So we are absolutely reinventing our hiring process to hire those “Renaissance technologists”. That moves away from either coding tests or “have you passed your Institute of Banking certificate?”, to be much more of a multi-dimensional interview and selection process. That’s a very different mindset too. And we’re investing very heavily in training and coaching our hiring managers to seek those people out. I think never has a growth mindset been more important than now, in this world—and particularly in financial services, when you think about risk management and all of those things. If you’re not able to adapt your thinking to a new paradigm, you can very easily be out of the game really quickly, and this technology is changing so fast. Being able to have a bit of courage, actually, as a leader—to adapt and to lead—and to be a bit raw and open about your own limitations, understanding, and learning journey is a really important part of leadership. But we role-model that and say, “Actually, we don’t have the answers either.” And doing this course, or experimenting with this tool—just to make that lifelong learning culture really visible—I think that’s really important.
Tom Rourke 11:18
And that role modeling, I presume, expands right across all of the disciplines in the bank and across all the domains of leadership, actually, fascinating. There are so many things we could go deep on, but again, Ralitsa, when you're out in the world, is that your experience as you're working with other organizations, or, like, how are the people beginning to emerge as you were following similar paths?
Ralitsa Nenkova 11:37
What I’m seeing with clients is that there are two buckets. There are those that have looked at AI literacy top-to-bottom and bottom-up, and there’s been a lot of training rationalization—how they’re adopting it across the organization, and the rest. And those are the ones that are really seeing the benefits, because it’s permeating through the culture and is part of the strategy. Then there are those where AI is still seen as a “technology thing”—something that technology wants to do—and they haven’t necessarily grasped that AI permeates the full fabric of the business. They have a way to go to get their hiring strategy, their talent, etc., right. To me, those two buckets will drive competitive advantage going forward. Those that move relatively rapidly into bucket one will really make a difference to the industry. And I think for those that don’t, and don’t invest in this—there’s a lot of consolidation in insurance currently, but we’ll see more of that because they won’t be able to survive the competitive pressures. Actually, a client said to me—and it’s a great quote—“AI is not a technology thing. This is not just the next technology, as when we were deploying software, or when the internet was invented. This is something that is becoming part of the fabric of the organization.” And those in bucket one that are really progressive: I’m hearing a lot more. Five different global insurers have said to me in the last month that they are monitoring and measuring their workforce now as the number of human employees as well as the number of agents—and how they’re looking after the agents and developing that IP as well, which is really where we’re all heading. And that, for me, is the fundamental mindset shift that everybody needs to embrace; because otherwise, if you rely continuously just on human power going forward, you will be left out of the race.
Diana Kennedy 13:35
One thing I like about NatWest is we have a really clear and simple set of goals: one, grow our customer base; two, simplify the bank; and three, manage our balance sheet better. And how we do that absolutely is enabled with AI. And I think if you talk to any employee anywhere in the bank, they will be able to say those things. And they would then say, “And how we do it—and our values—are that we think we win with customers, and we raise the bar.” And that culture of performance and raising the bar is really entrenched and pervasive in the organization, which is an amazing thing. However, when you’re adopting and embedding a new technology like this, getting the foundations right is super critical—particularly when you’re going faster. So we’re investing very heavily in things that perhaps you would focus on less if you were looking at this top-down: the controls and the behaviour of agents, and being able to manage that; to measure it; to view that at a dashboard level; and the identity of agents. So how do you manage the attestation of the action of an agent back to a human? What does that look like? And how do you make the registry of your agents? How are you going to put these things in place before they scale and become embedded in cupboards in the organization that you discover in 10 years’ time? So having control over how you register your agents—there’s a lot in there, from a foundational infrastructure point of view—to manage scaled AI adoption. But the value from the AI projects is exactly as before: productivity growth, balance sheet management, and simplification. Are we simplifying our processes, and can we measure those in the same way as we were previously—just that we now have a new, really fantastic tool in our kit bag that we didn’t have before, enabling us to go faster if we choose to. And you may not want to go faster; you may just want to be more controlled. You may just want to be more automated, because anything that frees up your people matters.
Tom Rourke 16:05
I like that comment about if you want to go faster because of something recently where, like, there are certain contexts where friction has a purpose and particularly in control. And I think that point you make about the visibility of the journey is super useful, because you know where I was going to ask next was around, like, lots of people tell their story. Of success at the end, and it looks fantastic, and everyone says, “Well, I want that, but the challenges were, where did you start?” Like, what are those first steps that leadership can take, or should be taking to kind of begin that journey?
Diana Kennedy 16:34
I do think winning the hearts and minds of colleagues is really important. So you're democratizing access to AI tooling, like your product, your personal productivity stack, in a safe way that you're encouraged to experiment and try new things, and there's this sort of scaffolding to enable you to do that. It's really, really important. And I think starting with colleagues is a really good place to start. Secondly, as you think about products, your product as a business, I think it is right to experiments more. I think one of the things that we try to do is, is to anchor everything back to customer. Of course, that's why we exist. But anchor any change or you feature our adaptation back to what the customer outcome is. I think where many organizations perhaps have failed or perhaps not achieved the progress that they've wanted is by just running endless POCs. I think it's very easy to run endless POCs. It's much harder to scale in an organization, and if you can take a POC, but make sure you're proving something that is then systemic within the organization, and then that's how you win. So that's perhaps where I'd start.
Ralitsa Nenkova 17:50
It's AI fluency across the organization and basically making people comfortable to welcome AI and not see it as a threat to their jobs. And the reality is things are moving so fast that you have to accept that you are transforming daily. And you have to have the mindset that this is your business. Transformation is ongoing and every day, and you have to continuously challenge yourself. You can't just accept that something that's been thought about a year ago, etc, can still be valid in 12 to 18 months. That's no longer the case, and I think that's another big mindset shift that we need to accept is part of our or put it another way, transformation is part of BAU. Now there's no more BAU in the old sense of the word, and this is where genetic comes in huge value, because the agents are giving you that access to the productivity and the ideas and the decision making that can happen almost instantly. That was never the case before.
Tom Rourke 18:56
Obviously, we're all partnering in this, you know, in this world is a lot more in choosing, but how when the pace is so fast, when you know you must look beyond capabilities, and it was interesting, you frequently come back to the word values, as you've been talking today. Who do you look for then, in terms of, I know you use the word mutuality when we were writing first recently, but like, what do you look for in a partner beyond capabilities and solutions in this context?
Diana Kennedy 19:21
Well, I would say mutuality—and partners and partnerships in times of transformation like this—are more important than ever. You choose partners not just on the capability that they bring; you choose them because they’re aligned to your values, mission, and purpose. And they have an ambition and set of goals that are aligned—where you can drive mutuality—where there’s value to both organizations, where one plus one can equal three. And we are very purposeful about that. Clearly, our partnership with Kyndryl is a very important one for that reason, because we see alignment on the mission and the goals that you have, and how you think about this transformation is very aligned to ours. But it’s also about having a strength of relationship where you can challenge each other—and that’s what partnership is about: having difficult conversations, where our partners can challenge us, and where we can challenge them back. And that’s really important, because high performance comes from being able to raise the bar through challenge and peer review—an environment where it’s safe to do that—because that’s how progress happens. Progress isn’t easy; it comes through safe, intentional, mutual challenge. And that’s what partnership is about.
Tom Rourke 20:52
As we’re coming to our close, I’m going to throw a challenge to both of you. Maybe we’ll start with you, Ralitsa. So we opened with the idea that this feels like the most exciting time to be in technology and to be in our industry, right? But if you look to 2026 and beyond, what would make it an even more exciting environment, in your view? What does progress look like for you?
Ralitsa Nenkova 21:14
I think for me, success for the industry overall would be that the customers are feeling the difference, and they're engaged a lot more with financial services. I've been in this industry for so long, and this is where we've always had a challenge, if you compare to other industries like retail, we just never could get people to get excited about their pension. And actually, you know, the concept of a fantastic retirement is close to everybody's heart, but they they were never engaged in the way that the industry tended to approach so if AI can help us to live and really genuinely delight the customers to a point where they're thinking about it, it's part of their lifestyle, and they really get excited about their financial wellbeing. I think this is a huge win. And I genuinely think AI is the first time where that is a genuine opportunity. The second part is that where we've really seen significant improvements to the bottom line in terms of commercial value, which then drives, obviously, the overall economy, because FSI is massive worldwide. And the third for me, in terms of success is that we have got everybody on the journey in terms of our workforce and talent, and we have finally, really genuinely our industry as FSI, where youngsters are keen and excited to join.
Diana Kennedy 22:40
I completely agree, perhaps, just perhaps two things. One, on the personal level, what 2026 looks like. So I'm encouraging my own children to experiment with AI, set on the task of using Claude codes to build an app that measures the vibration sensor on our tumble dryer, so they know when to go and empty the laundry. So in 22 years, it could be fantastic.
Tom Rourke 23:03
Having them empty the laundry.
Diana Kennedy 23:05
Exactly, exactly.
Diana Kennedy 23:07
So which leads to my second point, which is I am actually hugely optimistic about the opportunity and value of AI to not just my business, but to communities, to economies, to countries, to help them be lifted out of poverty. And I do genuinely believe that this technology is getting to the point of being so democratized that that's possible. And you know, if we can all leave this this world and have made it a better place. Then, you know, I think we should, we should feel very proud of that. And I think being able to lead our organization in a way that hopefully creates a bit of a shadow that things are possible and and if that inspires even one person to do more with this technology for their own community, then that would be an amazing thing.
Tom Rourke 24:05
Thank you so much for sharing that vision. Certainly it lifted my morning. And I can just say thank you both very much, Diana and Ralitsa, for joining us here on The Progress Report and giving of your time. Thank you.
Diana Kennedy 24:17
Thank you for having us.
Ralitsa Nenkova 24:19
It's been a real pleasure. Thank you. Bye.