EXCLUSIVE: “Building the GenAI Muscle” – Søren Andreasen, Nordea in ‘The Fintech Magazine’

Are Nordic institutions being overly cautious about genAI adoption or just waiting for the right moment to land a punch? We asked Nordea’s Søren Andreasen

Many anticipated that 2025 could be the year when AI gets over the hype cycle and begins to play a meaningful role in business. Seventy-seven per cent of bankers believe value derived from AI will be the differentiator between winners and losers in the coming years, according to The Economist. And there’s no doubt there’s serious investment behind it.

The International Monetary Fund says the financial sector’s AI spending is set to more than double by 2027. In practice, fintech strategists at 11:FS say that the breakthroughs will come from automating workflows, streamlining compliance, and enhancing fraud detection. But it has acknowledged there’s been a reality check of late as ‘banks discern [AI’s] true capabilities and limitations’.

According to recent research from Boston Consulting Group (BCG), only 26 per cent of companies have developed the capabilities to move beyond proofs of concept and generate tangible value. And there’s a noticeable change of pace around genAI specifically.

Take the Nordic countries, where you might expect enthusiastic adoption – they are, after all, renowned for being among the most comfortable with digital technology. And yet a survey of 110 business leaders across the Nordics by Cognizant and Oxford Economics revealed they were less optimistic about adopting genAI than elsewhere in Europe – their genAI ‘momentum score’ coming in 14 per cent below the average.

McKinsey had previously reported (in 2023) that AI, in all forms, was ‘not a priority on Nordic C-suites’ agenda’ – although that might have been because ‘the most significant barriers to adoption are a lack of talent, strategy, and technological infrastructure’. BCG was less empathetic, going so far as to say the Nordics suffer from ‘genAI complacency’.

Søren Andreasen who heads up Digital Customer Engagement at Nordea, the largest bank in the Nordics, with total assets of €584.7 billion, isn’t complacent, but he is cautious.

“It’s pretty evident that banks are laser-focused on genAI,” he says. “A lot are at a similar level, experimenting with it internally, or using it for a few customer applications.”

He told a FinTech Connect event in London in late 2024 that we may only see the transformational effect of the technology over the next five years. That doesn’t mean that Nordea is shying away from it, though.

“I put a lot of effort into making sure my team knows how to use genAI and what to use it for so, when they have a use case that makes sense, they can apply it,” says Andreasen.

The bank has even held hackathons to increase familiarity with the technology and is already using it internally, but as yet it’s not exposing genAI directly to customers.

“I put effort into making sure my teams know how to use genAI… so, when they have a use case that makes sense, they can apply it”

“It’s a way of building our organisational muscle, learning how to use genAI and discovering its pitfalls and challenges,” says Andreasen. “We know how to build models that check for biases, but it’s important that our customers’ data is kept safe. And, as a universal bank, we also know some of them are not really ready for genAI prompts yet.”

That said, the bank has implemented some initial use cases. Andreasen outlines just one.

“We are experimenting with generating marketing content, where our messages have to be personalised to many different languages, countries, customer segments and form formats,” he says. “Six languages x four countries x 20 customer segments x 10 channels (web, mobile, Facebook, LinkedIn, etc) will add up to hundreds of different expressions for each campaign we do.”

But, in general, he’s of the opinion that it’s still a nascent technology, and says it’s unclear where the best models will come from. The surprise launch of DeepSeek V3, which proved that models just as transformational as ChatGPT could be built for a fraction of the cost, demonstrated that. As Andreasen points out, billions may already have been wasted.

“Right now, it’s hard to see any significant business opportunities in the industry that could become a sustainable business model, which other banks can’t replicate,” he says. “The consensus is we’re probably five years away from really seeing the transformational aspects of genAI in financial services.”

Is there a danger, though, that hesitancy in the Nordic banking space will lead to complacency and loss of talent? Assessing the overall business landscape, BCG found that only 19 per cent of Nordic white-collar workers report using genAI on a weekly basis, compared to a global average of 61 per cent.

Andreasen is pragmatic. It’s not all or nothing, he says.

“Rather, with genAI, we’re moving towards a new norm, where this is just another tool in the banking toolbox”. As for business opportunities, ‘it will come as the technology matures’, says Andreasen. “And maybe, in five years’ time, we’ll have seen some really transformative use cases for it.” In theory, he agrees that there’s no better location to roll out the technology.

“Sweden, Norway, Denmark, and Finland are all super-digitised. Everybody has a digital ID, for instance. That gives us a great advantage when it comes to deploying technology and getting customers used to using our solutions.”

But with genAI, that’s still to come.

This article was published in The Fintech Magazine Issue 34, Page 45