How Will AI Affect Jobs? The UK Finance Industry Is Showing Us

For years, the debate over artificial intelligence and employment has been conducted largely in the abstract: forecasts, think-tank scenarios and warnings about what might happen to white-collar work. Britain’s financial services sector is turning that abstraction into something observable. Banks, insurers, asset managers and the professional services firms that orbit them have been among the earliest and most aggressive adopters of generative AI — and the effects on hiring, job design and career ladders are starting to show up in the real world.

Why finance went first

The City was always a likely testbed. Financial firms are rich in exactly the kind of work that large language models handle well: drafting documents, summarising research, writing and checking code, reconciling data, answering routine customer queries and producing first-pass compliance reviews. They also have the budgets to buy enterprise tools, the data to feed them, and relentless shareholder pressure to hold down costs.

Just as importantly, finance is heavily regulated and intensely measured. Headcount, cost-to-income ratios and productivity are reported quarterly. That makes the sector an unusually legible place to watch how a new technology filters through an organisation — and to see whether the promised savings actually materialise.

The shape of the change

The pattern emerging is less dramatic than mass redundancy announcements and more consequential than business as usual. Rather than firing large numbers of workers outright, many institutions appear to be slowing or reshaping recruitment: leaving roles unfilled when people leave, trimming the size of graduate and junior intakes, and redeploying staff into areas where human judgment, relationship-building or regulatory accountability still carry a premium.

That has particular implications for early-career workers. The traditional apprenticeship model in banking, accountancy and law-adjacent roles relied on juniors doing large volumes of routine work — building spreadsheets, formatting pitch decks, combing through filings — and learning the business by osmosis. If software now does much of that work in seconds, firms face an awkward question: how do you train the senior bankers and analysts of 2040 without the grunt work that used to teach them?

Augmentation, not just substitution

The other half of the story is job redesign. Many roles are not disappearing so much as being rewritten. Employees are expected to supervise AI output rather than produce it from scratch, which puts a premium on verification, domain expertise and the ability to spot when a confident-sounding answer is wrong. New functions are also appearing around model governance, data quality and AI risk — areas regulators are watching closely, since accountability for a flawed decision still rests with the firm, not the software.

That creates winners as well as losers. Workers who can combine financial knowledge with technical fluency are in demand. Those whose value lay mainly in processing volume are more exposed.

A preview for everyone else

Finance rarely stays a special case for long. Many of the tasks being automated in the City — drafting, summarising, coding, triaging customer requests — are common across law, consulting, media, marketing, public administration and corporate back offices. What is happening in UK financial services is therefore less an industry story than a leading indicator.

The open questions are familiar ones: whether productivity gains translate into growth and new roles or simply into smaller payrolls; whether displaced workers can retrain quickly enough; and whether the training pipeline for entry-level professionals can be rebuilt. Britain’s banks are not answering those questions deliberately. They are answering them in practice, one hiring decision at a time. Read More