SaaScada: Why UK Banks Are Struggling with AI

Banks remain trapped between ambition and execution when it comes to AI adoption, with new research suggesting the industry's foundations are too fragile to support the transformative technology many institutions claim to champion.

A survey of 150 UK banking IT leaders conducted by SaaScada reveals a sector confident in its potential yet struggling with basic implementation challenges.

While 81% of respondents believe AI has huge potential to transform financial services, the reality tells a different story about readiness.

The research reveals that 99% of banks are at least interested in exploring AI, yet they occupy vastly different stages of maturity.

Just 18% have reached an advanced stage where AI is fully embedded across their business operations, driving automation and decision-making with continuous investment.

A further 37% classify themselves as maturing, with AI solutions implemented in some areas and actively expanding use.

However, 32% remain in the developing phase, running proof-of-concept projects without full deployment, whilst 9% are still assessing AI's potential without having started pilots.

Another 3% express interest but cite regulatory uncertainty and funding constraints as barriers to beginning implementation.

This stands in stark contrast to recent industry headlines suggesting rapid adoption across the sector.

Research from IBM published earlier in the year claimed that generative AI implementation in banks had rocketed from 8% to 78% within a single year.

However, SaaScada's findings suggest this headline figure masks significant variation in how deeply AI has actually penetrated banking operations.

"There's a real danger that the gap between the can-dos and the can't-dos will become unbridgeable as AI takes hold," says Nelson Wootton, CEO and Co-Founder at SaaScada.

"Too many banks are still focused on window dressing and experimenting at the edges, rather than addressing the fundamentals."

Where banks are actually using AI

The research identifies clear patterns in AI deployment across the sector.

Customer-facing applications like automated savings tools have achieved 51% implementation, whilst fraud detection sits at 44%.

However, areas where AI could deliver substantial operational value remain largely unexplored.

Only 34% of banks have implemented AI-powered operational efficiency and compliance reporting systems, despite this representing one of the technology's most promising applications.

Credit scoring and loan approval, a function where AI could significantly streamline processes, has reached just 37% implementation.

Temenos research published in April 2025 found similar hesitancy, with 59% of banking executives citing concerns about AI hallucinations producing factually incorrect information.

"It's Groundhog Day," observes Steve Round, President and Co-Founder at SaaScada.

"Every time a new technology comes along, banks rush to bolt it onto the front end and call it transformation."

The data infrastructure problem

The research identifies data quality as the critical bottleneck preventing effective AI deployment.

Some 79% of respondents agree that without quality data foundations, banks will fail to keep up with AI-driven innovation.

A substantial 63% state that AI in finance is "going nowhere fast" without real-time data access.

The issue compounds when legacy systems enter the equation.

Two-thirds of IT leaders compare running AI on legacy core banking systems to fuelling an electric vehicle with petrol.

"Banks love to talk about being data-driven, but most still aren't built for the kind of real-time access that AI actually needs," explains Paul Payne, CTO at SaaScada.

"Without proper guardrails, structure and oversight, customer data can all too easily be exposed."

Regulatory uncertainty and fintech competition

Regulatory concerns present another significant barrier to adoption.

Some 68% of surveyed IT leaders report that regulatory uncertainty is slowing AI deployment, whilst 63% admit compliance requirements actively deter them from using the technology.

Despite these challenges, 67% believe stricter regulation represents a worthwhile trade-off for ensuring proper oversight.

Perhaps most concerning for traditional banks is the competitive landscape.

Some 79% of respondents acknowledge that fintech challengers are racing ahead whilst established institutions remain bogged down by outdated infrastructure and complex processes.

Nelson connects this competitive disadvantage directly to infrastructure choices.

"Fintechs and challengers are embedding AI deep into their business models, using it to drive real value," he notes.

"But ultimately, it all comes back to the core."

Requirements for successful AI adoption

The research identifies three critical factors for successful AI implementation in financial services.

Clear regulatory guidelines and data privacy alignment top the list, followed by enhanced data quality and accessibility requiring standardised, real-time and reliable information.

The third factor, seamless integration with core banking platforms, represents the fundamental challenge many institutions have yet to address.

"The banks that invest in a modern core now will be the ones leading the AI revolution tomorrow," argues Steve.

"The rest will be left trying to retrofit the future onto the past and wondering why it doesn't work."