03 July 2026
- In our view, European banks are emerging as some of the early earnings beneficiaries of AI.
- The winners could be banks that turn productivity gains into lower costs and higher returns.
- Strong fundamentals and reasonable valuations can create investor appeal.
Since OpenAI unveiled ChatGPT at the end of November 20221, investors have poured money into the companies building AI. Now they are starting to ask a different question: who is actually making money from it?
One answer may be hiding in plain sight. In our view, European banks are emerging as some of the early beneficiaries of the AI boom. Unlike many industries still experimenting with the technology, banks are already reporting faster processing, lower headcount, better customer service and, crucially, improving financial results.
We believe that matters because banks are exactly the sort of organisations AI should help most. They employ thousands of software developers, operate vast back-office functions and spend billions every year processing information, handling customer enquiries and managing risk. If AI can genuinely automate knowledge work, banking is one of the first places investors could expect to see it.
The implications extend beyond the sector itself. The market's AI winners have so far been concentrated among semiconductor companies and hyperscalers. But if productivity gains continue to show up in bank earnings, we think the market may start looking beyond the companies selling AI and towards the companies using it. European banks, many still trading on modest valuations despite improving returns, could be among the first beneficiaries of that shift.
AI is starting to shrink the workforce
Bank headcount has been falling for years. What feels different this time is why.
This is no longer about closing branches or cleaning up after a crisis. Increasingly, banks are using AI to automate work that previously required large numbers of people. Management teams are becoming more explicit about what that means for staffing levels.
BNP Paribas has committed to reducing the workforce in its French retail banking division by 2.2–2.5% annually through 20302 Intesa Sanpaolo announced plans to cut over 6,000 jobs by 20293. ING has warned that up to 950 jobs in the Netherlands could be at risk by the end of 20264. Nordea CFO Ian Smith was even more direct: "We will see the number of full-time employees coming down over the next few years. There's no question about that."5
The evidence is showing up in day-to-day operations. NatWest says improvements to its AI-powered assistant Cora increased query resolution rates by 20 percentage points, helping drive the retail bank's cost-income ratio from 50% to 45% in a single year6. At Danske Bank, 95% of developers are now using generative AI coding tools, corporate credit processes are running 40% faster and roughly three-quarters of customer enquiries are resolved without human intervention.7
The same story is emerging in the United States. JPMorgan Chase's Marianne Lake told investors at the bank's 2025 Investor Day that operations headcount is expected to fall approximately 10% over the next five years - even as the business grows by more than 25%.8
The savings are starting to show up in earnings...
Santander has set a target for AI to generate more than EUR 1 billion in cumulative business value between 2026 and 2028 through a combination of higher revenues and lower costs.9 Management expects AI initiatives to contribute around one percentage point to the group's cost-income improvement between 2026 and 2028.10
Source: Santander Investor Day presentation, February 2026.
Commerzbank recently raised its 2028 return-on-tangible-equity target to 17%, citing greater use of AI as part of a strategy to drive efficiency.11 Lloyds Bank reported GBP 50 million of value from generative AI in 2025, and expects that figure to exceed GBP 100 million this year.12 NatWest delivered more than GBP 100 m of additional cost savings in the first quarter of 2026 alone, supported by productivity gains and the scaling of AI.13
The American banks provide a glimpse of where this could ultimately lead. JPMorgan estimates it is already generating around USD 2 billion annually of benefits from AI across applications including fraud prevention, trading, customer personalisation and operational efficiency.14 Goldman Sachs has deployed AI assistants across the firm and is testing autonomous software-development agents.15 Morgan Stanley sees AI-driven personalisation as a major competitive advantage in wealth management.16
...But extracting value must be targeted
In March, while recognising the measurable savings case, UniCredit CEO Andrea Orcel warned against indiscriminate rollout of AI tools, arguing that benefits depend on targeted deployment and measurable returns rather than blanket adoption.17
As banks move from pilot programmes to enterprise-wide rollouts, the economics become more complicated. Unlike traditional software licences, generative AI comes with an ongoing usage bill: every query, prompt and agent consumes computing power and tokens. The risk is that enthusiasm runs ahead of discipline.
ING's COO, Marnix van Stiphout, also favours a targeting approach, saying, "We're not gonna throw agentic AI at everything. It's gonna be very expensive if that's what we do."18
The bigger challenge is turning productivity gains into shareholder returns. Faster coding, quicker loan processing and more efficient customer service only matter if they ultimately reduce costs or increase revenues. The upfront spending on infrastructure, data and implementation arrives long before the benefits. In our view, banks that fail to translate productivity gains into lower headcount or leaner operations risk simply layering AI costs onto an already complex cost base, potentially diluting returns in the meantime.
There are reputational risks too. Standard Chartered recently faced public criticism after an AI-generated client communication contained factual inaccuracies, a reminder that customer-facing AI mistakes can quickly become headline news.19 In banking, where trust is the product, the margin for error is slim.
Source: Financial Times, 21 May 2026
Why European banks appear attractive today
We think that the AI story is arriving at a good time for European banks. The sector's fundamentals were already improving before AI entered the debate, and recent geopolitical volatility has improved valuations, enhancing the sector’s appeal for potential investors. The Middle East-driven sell-off reset valuations that had become increasingly stretched after a multi-year rally, even as earnings expectations have continued to improve.
The backdrop remains supportive, in our view; balance sheets are stronger than they were a decade ago, much of the riskier lending has migrated into private credit markets, and gradual European Central Bank rate cuts mean replication portfolios should continue to support earnings for years to come.
Source: GAM, Bloomberg, June 2026
Within the sector, banks with exposure to faster-growing parts of Europe appear well positioned, particularly Greece, Spain, Ireland and Eastern Europe, where stronger economic growth, rising credit demand and EU investment flows should provide additional support.
Tom O’Hara, David Barker and Jamie Ross manage European Equities strategies at GAM Investments. You can find out more information on the team here.