Videos
Banking on AI: Carsten Egeriis on Infrastructure, Adoption, and Real Return
Insights
- Banking has always been among the heaviest technology investors of any sector, and AI is now accelerating work already underway across cloud migration, data analytics, and customer journeys.
- Some AI applications are easy to measure, like chatbot resolution rates in the mobile bank. Others, like agentic credit processes, are still too early stage to yield a clear return on investment figure.
- Open-source tools have closed the gap in areas Danske Bank once saw as a competitive edge, prompting a shift toward heavier investment in infrastructure and data products rather than applications.
Carsten Egeriis, CEO of Danske Bank, describes an organization moving at different speeds on AI. Tools like Copilot and automated meeting summaries are already part of daily work, while larger initiatives, such as end-to-end credit approval or know-your-customer processes, are still being built piece by piece rather than as fully automated flows. Carsten attributes the gap between what AI can do and what the bank uses it for to culture, adoption, and the slower pace of regulatory approval for generative models compared to traditional banking systems. He points to Infosys as a partner that gives Danske Bank access to best practices and tooling insight it couldn't develop as quickly alone. Rather than treating AI as a single strategic bet, Carsten frames it as a sustained investment across talent, infrastructure, and modernization, something he sees as a requirement rather than a gamble.
Carsten Egeriis:
There are certain areas of AI where I thought we were building competitive advantage, but where it's clear to me there is no competitive advantage because the tools are available open source in the market, and you can use it pretty fast. So I would've doubled down even more on building infrastructure and building data products, and less so on the application side of things.
Carsten Egeriis:
Hi, I'm Carsten Egeriis. I'm CEO of Danske Bank.
Carsten Egeriis:
So if you look at banking, banking has already been and always been huge investors in technology and probably the sector that invests the most in technology across all sectors globally. So technology is incredibly important for us, and keeping pace is incredibly important. And the way I think about AI is that it sort of can augment and supercharge everything we do. So we've already been on a path to move applications on cloud, to modernize our technology, to do data analytics, modeling, and AI is a way to sort of supercharge that across the entire processes, customer journeys in banking.
Carsten Egeriis:
So if I look at AI and how it's actually been embedded, implemented, within the bank, there are some areas where I think it's more or less embedded, and those areas would be things like using tools for software development, for using tools, for example, Copilot, to do meeting minutes with our customers. So you automate meeting minutes. You don't have to do it manually. You can send it out to customers. You can be more present with your client. So those are small examples where you can just use the tools for automating and for accelerating work. Then there are other areas where we're looking at end-to-end customer journeys, like a credit process or a know your customer journey process or preparing for client interaction work. And there we're sort of in the process of building the agents, but we're not at a place where it's a full agentic flow where the agents speak with each other and you have a fully automated and fully sort of AI-proof process. So we're in different sort of areas of development depending on sort of which area of the bank we're talking about.
Carsten Egeriis:
So I think measuring whether AI value is real and what the return on investment is on your AI investments is probably the most asked question, both from my board and from investors. Look, I think there are different ways of doing it, and there are some that are very straightforward. So, for example, we've built an AI chatbot in our mobile bank, and there we can measure exactly how many customer queries are answered by the AI chatbot. So there it's very easy to both measure customer engagement, customer satisfaction, but also how many customer queries you resolve. And then there are of course areas where it's much more difficult to measure value, and that is such as building agentic flows for a credit process. It's not automated yet, it's only a very few parts of our credit process that uses AI today, for example pre-populated customer profile. So it's more difficult at this stage to look at what is the true return on investment of the AI. And it's also not clear to us what is the best approach yet. But I think overall, there is no question that the return on investment on AI is going to be very significant. And if you're not at the cutting edge and if you're not testing and working with AI, then you're going to be left behind. And therefore, I think we really need to sometimes challenge each other when we think about what are returns, what metrics do we measure when we think about this. And there, I think you need to be much more nuanced, again, based on what area of the business is it that you are embedding and rolling AI out in.
Carsten Egeriis:
I think the interesting thing with AI is that the technology is there and can do so much more than what we're actually using the technology for. And therefore, it is partly around adoption and culture, and we're driving that hard, and we can monitor adoption, and we can do hackathons and AI weeks and communication and demos of what we're using AI for in the organization. So that's one part of it. But the other part of it is there is no question that in banking we're a regulated sector. And so there is also this catch-up from a regulatory perspective around actually being able to, for example, get models approved. And these are very different models than your traditional banking models when you talk generative AI type models. So it's different things. It's adoption, it's culture, it's risk management, it's getting things embedded in processes, it's getting things out to the customer and how you test and learn around that. So I think there is many different areas that create this capability overhang. The technology is there, but it's not truly embedded.
Carsten Egeriis:
So Infosys is a very significant partner for us. It's one of our absolute biggest partners. And we have a significant amount of our software engineering, software skills, project management skills, and actually our Infosys team is very well embedded and integrated into a lot of our agile ways of working. So basically, everything we're working on, whether it's our mobile bank, whether it's our corporate banking model, whether it's how we think about accelerating our product development life cycle, Infosys is involved in that. So it's an incredibly important partner, and what it gives us is that, look, although we are big investors in technology, and although we spend a lot of time on technology, there's no way that we can get the amount of best practice sharing, the amount of testing with different tools, or the amount of knowledge and insight into the development of what's happening in AI without a partner like Infosys. So for us, it's really about how do we partner up the best of banking and the best of technology to create better and more value for our clients.
Carsten Egeriis:
There are certain areas of AI where I thought we were building competitive advantage, but where it's clear to me there is no competitive advantage because the tools are available open source in the market. And you can use it pretty fast. So I would've doubled down even more on building infrastructure and building data products, and less so on the application side of things. If there was one sort of, I guess, bigger bet that we're making, I would generalize it because there's not one big bet, but across the board, we're really investing heavily in augmenting and accelerating our AI journey across pretty much all our customer journeys. So it's infrastructure, it's modernization of our tech stack, it's making sure that we invest heavily in talent and training. And I don't really see it as a bet. I see it as a must.
