Outdated IT & regulation slow UK banks’ drive for AI adoption

New data shows that UK banks are adopting artificial intelligence (AI) at a slower rate than many of their leaders expect, with outdated systems, data challenges, and regulatory uncertainty hindering further growth in the sector.

AI optimism

Research surveying 150 UK banking leaders found that 80% believe the country's financial sector is well positioned to take advantage of AI. However, only 55% of the banks surveyed have so far deployed AI within their organisations. This gap highlights the continued challenges facing banks as excitement around AI adoption outpaces practical implementation.

Despite major investment across finance, a substantial proportion of UK banks are yet to move beyond the pilot stage. The survey also reveals that 81% of leaders think AI will have a significant impact on the industry. However, the timing of this remains under debate: while 32% say AI is already having a transformative effect and 28% expect it within the year, a notable 21% believe industry-wide impact is still years away.

Legacy constraints

The research found major barriers related to technology infrastructure. Outdated legacy technology remains a significant drag on progress. Of those surveyed, 66% likened the challenge of running AI on legacy systems to "fuelling an EV with petrol," and 63% said real-time access to high quality, accurate data is lacking. 79% agreed that data foundations are crucial for sustained AI innovation.

"AI isn't magic. It's maths, data, and timing. If your systems can't deliver the data when and where it's needed, no amount of clever tech on top will fix it," said Steve Round, President and Co-Founder, SaaScada. "The banks that invest in a modern core now will be the ones leading the AI revolution tomorrow. Ultimately, you can't build the future on foundations from the past."

The allure of new technology is driving many banks to invest mainly in lower-risk AI use cases at the customer interface such as automated savings and bill management. The most widely implemented application is AI-powered savings and smart money management, with 51% deployment. AI-powered fraud detection, subscription management, and cashflow forecasting are each implemented by 44% of respondents. More complex uses, such as operational efficiency or compliance reporting, saw uptake around 34%.

Regulatory uncertainty

Regulatory clarity is another key concern. 68% of respondents say uncertainty over regulatory requirements is causing banks to hesitate on AI adoption. Additionally, 63% say that increased compliance and reporting burdens are discouraging wider deployment.

While most acknowledge the need for regulation, there is little appetite for a fundamentally new regulatory regime targeting AI. Instead, 67% favour adapting current financial regulations, feeling that oversight is essential to building trust and preventing misuse, even if it causes some short-term slowdown. 54% believe the Financial Conduct Authority's approach will address key risks such as bias, data inaccuracy, and privacy.

"The FCA isn't going to reinvent the rulebook for AI and nor should it. It'll fold AI into existing principles and judge firms on outcomes. That's the right approach," said Nelson Wootton, Co-Founder and CEO, SaaScada. "The guardrails already exist, and waiting for new rules is just an excuse not to act. Banks burying their heads in the sand are missing the point - AI won't just need compliance, it'll assist with compliance. But only if they get their act together on data."

Risk concerns

Concerns about AI risk centre around three main areas: data security, fears that overreliance will reduce skills within banks, and the potential for discriminatory outcomes from biased data. These issues have resulted in most banks keeping their AI use focused on automation and support functions where risks can be closely managed.

For successful adoption, respondents highlighted three priorities: clear regulatory guidelines, improved data quality and accessibility, and seamless integration of AI with core banking platforms.

"It's Groundhog Day. A new technology comes along, and banks rush to bolt it to the front end, calling it transformation. But until they do the hard work of modernising the core, it's just lipstick on legacy," said Wootton. "If your plumbing's broken, painting the door red won't fix the leaky tap. The real benefits of AI will come when it's supported by a modern, cloud-native, API-driven core. With the right architecture, data can move in real time, systems can talk to each other, and decisions can be made in the moment - not weeks later - giving banks the perfect platform to start to see tangible AI success."