Three-quarters of UK financial services firms are already using AI, up from 58% just two years earlier. Yet, only a third of those firms say they fully understand the systems they’ve deployed. The gap between adoption and comprehension is a clear signal that AI within financial services has moved past the question of whether firms will adopt the technology, and is now whether transformation functions can be resourced to safely deliver inside a regulated environment.
Ambition isn’t a constraint for AI transformation. Boards have approved the budgets, procurement has signed the vendor contracts, and strategy documents across banking, insurance and asset management now reference AI as a given. What hasn’t scaled at the same rate is the pool of people who can design, govern and land AI-enabled change inside a live transformation programme.
Where AI is landing
The Bank of England’s survey found the greatest perceived benefits of AI are in data and analytical insight, anti-money laundering and fraud detection, and cybersecurity. Customer support, compliance and reporting are where firms are expecting the next wave of AI use cases to focus on. The data shows AI isn’t confined to innovation labs or pilot projects, but embedded into what keeps a business running. The challenge now isn’t whether businesses can experiment with AI, but whether their existing delivery teams can govern it properly once it’s live.
The reason this gap exists is structural rather than a reflection of a businesses lack of will. Firms are increasingly leaning on third parties to build what they deploy: a third of all AI use cases identified in the 2024 survey were third-party implementations, nearly double the share recorded in 2022. Although this shift brings real advantages in speed, it also means a growing share of firms are accountable for models they didn’t build and may not fully understand, at a moment when only 34% report complete understanding of the AI systems already in use.
Layer regulatory expectation on top of this, and it becomes evident that a transformation lead who can manage a programme timeline is no longer sufficient. The PRA’s model risk management framework and obligations around data protection, operational resilience, cyber and Consumer Duty all raise the bar for who is qualified to own an AI system inside a governance structure. The role has now evolved to require someone who can also sit in a model risk committee and defend a decision to a regulator.
The AI skills shortfall
The scale of the skills shortfall behind this is easy to underestimate because it doesn’t look like a conventional talent gap. The Financial Services Skills Commission puts genuine AI specialists at around 1.5% of the sector’s workforce, alongside a 35 percentage point gap between demand for AI-related skills and the talent available to meet it.
The FSSC’s report published this year estimates the sector will need to recruit and retrain 450,000 people over the next decade to keep pace with AI and other disruptive technologies. The pattern that emerges is a thin layer of people who have delivered AI transformation at scale, a much larger group of technically capable engineers and change professionals who haven’t yet run a programme like this, and a workforce-wide need for baseline AI literacy that sits below both.
Most transformation functions are built to operate with exactly that shape: a handful of senior leads, a broader delivery layer, and wide literacy requirements underneath. The problem is that firms are competing hardest for the thinnest layer, the one with almost nothing to draw from because the discipline itself barely existed five years ago.
Shortages that aren’t matching the job spec
For anyone building a transformation team, the practical difficulty is that the profiles in shortest supply rarely match a standard job specification. Roles like an AI-literate transformation lead, or a model risk specialist who understands both the mathematics and the regulatory rulebook all sit between functions rather than inside one, and job titles across the market reflect this inconsistency.
A role that would once have been a straightforward Change Manager brief now often needs to absorb elements of governance, technical fluency and stakeholder management that used to sit in separate hires. The instability in how roles are titled and scoped is itself evidence that the shortage is being underestimated. Firms searching against a generic “AI Engineer” or “Transformation Manager” template are searching in the wrong place.
The AI hiring market today
The hiring market is already responding to this pressure, even where the job titles haven’t caught up. Morgan McKinley’s London Employment Monitor recorded a 12% year-on-year rise in UK financial services vacancies in 2025, with AI, data and regulatory roles cited as the main driver, and risk and compliance vacancies within fintech projected to rise 26% year-on-year over the same period.
KPMG’s most recent sentiment survey of UK financial leaders found 52% of firms hiring in 2026 expect that recruitment will concentrate on technology roles, with AI skills the single biggest priority for both external hiring and internal upskilling, and 57% of firms expanding at board level named AI expertise as their top acquisition target.
Vacancy growth and job-title fragmentation are showing the demand for a hybrid profile, even as the market struggles to describe what that profile is.
Finding the people who fit
None of this is a case for waiting until the talent pool matures. The next phase of AI-enabled transformation in UK financial services won’t be limited by budget or by strategic intent. Both of those are already in place. It will be limited by whether firms can build transformation teams with people who understand the models they’re deploying and the regulatory framework those models sit inside.
JSS spends time in exactly that thin layer of the market, working with candidates whose CVs don’t always say “AI” but whose track record proves they’ve delivered it inside a regulated environment. If you’re building out a transformation function and struggling to find profiles that match what the role needs, talk to us.
