PwC’s 2026 Financial Services Workforce AI Survey found something worth sitting with: only 42% of financial services leaders say they’ve actually done the high-level modelling to understand how AI is changing labour capacity across their organisation.
The rest are making workforce decisions, hiring plans, restructuring, where to invest in skills, without that evidence underneath them.
That’s the real gap. Not whether AI will change headcount somewhere, which most leaders already expect to some degree, but whether the plan behind that expectation is built on evidence or on assumption. For most firms, right now, it’s an assumption, and that’s a much shakier foundation for a workforce strategy than the claimed confidence in most boardrooms would suggest.
The reason that gap exists isn’t a lack of ambition. It’s that most AI workforce planning starts from the org chart and the job description – what a role is supposed to involve on paper – rather than from evidence of where the work is actually happening day to day.
Job descriptions age quickly and were rarely accurate to begin with. Process maps get drawn once and rarely revisited. Neither tells you, with any confidence, where a genuine AI opportunity sits versus where it’s simply the loudest idea in the room.
We’ve watched that dynamic play out directly. At our AI Capability Building session in Belfast, participants described a structural problem sitting underneath most AI workforce plans: budget increasingly flows almost exclusively to projects labelled “AI.” The predictable result is teams retrofitting an AI angle onto work that never needed one, simply to access funding. Deployments that nobody asked for, governed by nobody who understands them, solving problems that didn’t exist. That’s not a failure of individual judgement. It’s what happens when initiatives chased opportunistically instead of measured against where value and readiness actually intersect.
PwC’s own 2026 Global AI Jobs Barometer, drawn from over a billion job advertisements worldwide, found the labour market splitting into two distinct tracks: roles where AI automates the routine work so human judgement and expertise become more valuable, and roles where AI simply makes the job easier for anyone to do. Which track a given role falls into isn’t obvious from its title. It has to be discovered, measured against the real problem it solves and the value it protects, not assumed from the org chart.
The alternative isn’t more caution. It’s a different starting point: measuring where work genuinely gets done before deciding where AI goes next, so opportunities can be ranked by value and delivery readiness rather than pursued because a team happened to raise its hand first. That ranking needs a named owner responsible for turning it into an actual roadmap, because opportunity without accountability is exactly how well-intentioned pilots quietly stall in the “permanent pilot” stage so much of the market is currently stuck in.
The workforce development side of this only works once that evidence exists. People can’t meaningfully shift toward higher-value, judgement-heavy work, the side of PwC’s two-track split that’s actually growing, if nobody’s told them which challenges are the organisation’s real priorities. Generic upskilling, aimed at everyone equally, is a much weaker lever than development targeted at the specific gaps the evidence reveals. And because AI adoption doesn’t hold still, that picture needs refreshing on a real cycle, not set once and left to age the same way the original job descriptions did.
Our free 15-minute Leading through Disruption diagnostic, developed in partnership with SolasOS includes a dedicated section on AI adoption and workforce development – an instant, evidence-based read on whether your workforce plan is built on real signal, or on assumption.
Start the diagnostic: solas-os.ai/readiness
This closes our five-part series on the pillars behind Leading through Disruption. If you haven’t taken it yet, the full 15-minute diagnostic covers all four areas – leadership, talent, technology teams and workforce transformation – in one sitting.
This article draws on discussion from FinTech-Tables’ AI Capability Building session, AI & Data Leadership Forum (Belfast, June 2026), alongside PwC’s 2026 Financial Services Workforce AI Survey and PwC’s 2026 Global AI Jobs Barometer.
