4 min read
The CFO in the AI Era — UK Banking & Financial Services
Stop Hiding Behind the Pilot
UK banking has already adopted AI. Most CFOs just haven’t noticed it’s their job now.
Three in four UK financial services firms already use AI. That fight is over. The fight most CFOs are quietly losing is what happens to finance leadership once the transactional work that used to justify it is gone. IBM’s Institute for Business Value is blunt: only a third of finance functions are genuinely optimising the AI they’ve deployed. The rest are running pilots and hoping nobody asks what the close-cycle KPI is still for.
The comfortable line — every vendor deck repeats it — is that AI “frees finance for higher-value work.” Fine. But almost nobody has rewritten the scorecard to check whether that’s actually happening, or whether the CFO is just presiding over a smaller, faster version of the same reporting function. Three changes UK banking CFOs can’t keep deferring, below.
A Leadership Problem, Not a Staffing Slide
Finance has always built judgment the slow way: a controller’s nose for a misstatement, a treasury analyst’s gut feel for a liquidity squeeze. That instinct was earned doing the transactional grind — reconciliations, variance packs, first-pass reporting — that AI is now quietly eating. If the apprenticeship disappears before the next generation’s judgment is built, banks aren’t saving cost. They’re eating their own succession pipeline and calling it efficiency.
IBM’s 2026 CEO Study says the polite version of the same thing: organisations that redesigned five core business areas, finance included, were four times more likely to deliver on their objectives. Redesigned. Not automated. Three shifts define the real job now:
- Augmentation at the top, not just the bottom. Planning cycles are compressing from weeks to hours — that only matters if the freed-up time goes to judgment, not more meetings about the AI roadmap.
- An accountability gap most CFOs would rather not audit. Most finance leaders say the CFO should own AI outcomes in finance; far fewer have actually assigned that ownership cleanly. In a UK bank, that vagueness doesn’t survive contact with SMCR.
- A talent pipeline quietly being strip-mined. IBM finds 77% of organisations now see talent and technology leadership converging — tighter than most finance functions are staffed or trained for.
Why UK Banking Doesn’t Get to Sit This Out
Generic “CFO and AI” thought leadership misses the fact that actually matters here: in banking, finance output feeds regulatory returns, capital adequacy, and credit decisions. You don’t get to experiment your way through that.
The PRA’s Supervisory Statement SS1/23 — seven principles covering model governance, validation, deployment, decommissioning — runs alongside the FCA’s Discussion Paper DP5/22, which applies to any FCA-regulated firm using AI materially, no size threshold, no “we’re still piloting” exemption. Neither regulator has written AI-specific rules; both have told the Treasury Committee the existing framework is sufficient for now — while reserving the right to act on ‘egregious failures.’ Read calmly, that sounds reasonable. Read as a CFO, it should sound like regulators betting you’ll govern yourself properly without being told exactly how.
Under SMCR, a Senior Management Function holder carries overall responsibility for every business area — including AI controls inside finance — even when SMF24 or SMF4 nominally owns the model or the risk framework. “IT built it” is not a defence. The Treasury Committee has pushed the FCA to publish, by end-2026, guidance on senior manager accountability for AI-caused harm. That’s a deadline with your name on it. Consumer Duty adds a second UK-specific exposure: a biased affordability check or misleading auto-generated letter is a breach whether or not anyone meant it to be — and finance-originated AI increasingly touches pricing, affordability, and collections.
The Line That Should Be on Every CFO’s Wall
Your job used to be certifying the numbers were right. It’s now certifying that the system that produced them — human and machine — can be explained and defended under SMCR. If you can’t do that in front of an examiner, the AI strategy isn’t a strategy. It’s an exposure.
The KPIs That Actually Tell You Something
Close-cycle time and forecast accuracy aren’t wrong — they’re just measuring a shrinking layer of the business. These three groups tell a board or a regulator whether the CFO is building what matters next.
1. Judgment & Strategic Capacity
| KPI | What it measures | Replaces | Target signal |
|---|---|---|---|
| Strategic time ratio | % of senior finance time on capital allocation and risk appetite vs. transactional oversight | Time-spent surveys measuring activity, not value | Sustained increase, not a one-quarter spike |
| AI-assisted scenario throughput | Strategic scenarios modelled per quarter beyond the mandatory PRA stress-test calendar | Annual planning as a compliance exercise | A real increase — finance shaping strategy, not narrating it |
2. Governance & Trust — the one most CFOs will fail first
| KPI | What it measures | Replaces | Target signal |
|---|---|---|---|
| SS1/23 model inventory coverage | % of AI/ML tools in finance and credit-adjacent decisions inventoried and validated under SS1/23 | Shadow AI nobody admits to in the steering pack | Approaching 100%, exam-ready — not assembled the week before |
| SMF accountability clarity | Whether AI ownership for finance-originated tools is documented against a named Senior Management Function | “We assume that’s covered elsewhere” | Documented, statement-of-responsibilities-ready |
3. Talent & Capability Renewal
| KPI | What it measures | Replaces | Target signal |
|---|---|---|---|
| Judgment-building pathway redesign | A deliberate programme — rotations, case-based review, mentorship — replacing transactional work as the training ground | Hoping junior staff “pick it up” by osmosis | A documented pathway people actually go through |
| Talent-technology role convergence | % of finance leadership roles with AI/technology fluency built into the actual mandate | Treating AI as a separate, CIO-owned skill | Rising, in line with IBM’s 77% convergence finding |
The Closing Argument
UK financial services has already answered the adoption question. What hasn’t been answered is whether finance leadership changed alongside it, or just got a faster version of the old job. SS1/23, SMCR, and Consumer Duty mean that gap isn’t a strategic nicety — it’s an exam finding waiting to happen. The banks that come out ahead will be the ones whose CFOs rewrote what finance leadership measures before a regulator made them.
Sources
- IBM Institute for Business Value, 2026 Global Outlook for Banking and Financial Markets: ibm.com
- IBM, FP&A 2026 Trends (69% of CFOs say AI is integral to finance transformation): ibm.com
- IBM Newsroom, 2026 IBM IBV CEO Study — CEOs Reshaping C-Suite Roles for the AI Era: newsroom.ibm.com
- UK Parliament Treasury Committee, Artificial Intelligence in Financial Services (HC 684): publications.parliament.uk
- UK Parliament Treasury Committee, regulator responses to the AI in Financial Services report: publications.parliament.uk
- BCLP, AI Regulation in Financial Services: Turning Principles into Practice: bclplaw.com
- FluxForce, FCA AI Paper: Requirements, Who It Applies To & Penalties: fluxforce.ai
- FCA, AI and the FCA: Our Approach: fca.org.uk