by Claude Opus 5.5

Which UK sectors appear most likely to see near-term AI-driven task change (e.g., professional services, finance, retail, public sector, healthcare), which have shown measurable change in 2026, and why?

The sectors where AI is most likely to change tasks soon are the same as in January: finance, professional services, IT, customer contact and public administration. All of them run on text, rules and screens. What is new in 2026 is that some change can now be measured, but almost all of it shows up in hiring (fewer vacancies, fewer entry-level posts) rather than in redundancies. In most sectors the evidence is still too thin to say how much of the fall is AI rather than costs and a weak economy.

The pattern behind the list

Near-term task change needs three things together. The work has to be digitised language or data. The tools have to be cheap to deploy on top of existing systems. And the organisation has to be allowed to rely on the output. Finance and professional services meet all three. The NHS meets the first two for paperwork but faces heavy permission barriers. Retail meets them in head office and contact centres, not on the shop floor.

Labour costs give firms a fourth push. Employer NICs, the National Living Wage and the Employment Rights Act have raised the cost of a marginal hire. Several surveys find firms meeting demand “by investing in automation and AI, rather than by raising headcount”, as Bank of England Agents put it in February 2026. That makes it harder to separate an AI effect from a cost effect, because both show up as hiring that didn’t happen.

Sector by sector

  • IT and software. Expected near-term task change: Code generation, testing, documentation, data work. Measured change in 2026: Entry-level software engineer hiring −27% y/y and data analysts −15% (DSIT/LinkedIn, Apr 2026). The BoE says software/IT consulting added 0.1pp a year to productivity growth in 2023–25, ten times its pre-Covid rate (reported). 58% of information and communication firms use AI (ONS, Jul 2026). Evidence quality: Moderate to good. Productivity, adoption and hiring data all point the same way.

  • Customer contact and admin. Expected near-term task change: Chatbots, agent assist, call summaries, triage, form-filling. Measured change in 2026: Vacancies in customer service and admin roles fell more than 20% over three years (Bank Underground staff blog, Aug 2026). British Gas is cutting 1,300 call-centre roles but says “AI isn’t driving these particular job reductions”; unions dispute that. Evidence quality: Moderate. The vacancy fall is clear, but the strongest number comes from staff research, not Bank policy, and firms deny the AI link.

  • Professional services (accounting, law, consulting). Expected near-term task change: Drafting, review, research, audit testing, reconciliations. Measured change in 2026: Entry-level accountant hiring −29% and legal assistants −14% (DSIT/LinkedIn). Big Four combined graduate and apprentice intake fell from 6,500 (2023) to 5,400 (2025). Trainee solicitor contracts −1.7% across 100+ firms (2026). Evidence quality: Moderate for entry-level hiring. Weak on cause: Legal Cheek mainly blames the shift to solicitor apprenticeships.

  • Finance and insurance. Expected near-term task change: KYC/AML review, claims triage, fraud, compliance drafting, analyst work. Measured change in 2026: London finance-analyst vacancies about 80, against 350+ four years earlier (Bloomberg, Jun 2026; unnamed job board). Standard Chartered plans about 7,800 back-office cuts by 2030 (reported; not clear how many are in the UK), and HSBC is reported to be weighing about 20,000. Evidence quality: Weak to moderate. Firm announcements are explicit about AI, but there are no systematic sector data.

  • Public administration. Expected near-term task change: Correspondence, consultation analysis, minutes, casework summaries. Measured change in 2026: Civil service headcount 557,000 in June 2026: down 1,000 on the quarter (the first quarterly fall since 2021) but up 6,000 on the year. Tools are deployed (HMRC: 32,000 Copilot licences, rising to 50,000). Evidence quality: Good on headcount, weak on attribution. There is no AI headcount target, and the falls have come from exit schemes.

  • Healthcare. Expected near-term task change: Clinical documentation, letters, coding, imaging support. Measured change in 2026: NHS employment 2.07m in June 2026, flat. Ambient voice tools are live across the Midlands (70,000 clinicians covered). A third of chest X-rays, about 2.4m, are AI-assisted. Evidence quality: Good on headcount and deployment. Time savings are measured locally, with no workforce effect.

  • Retail. Expected near-term task change: Head-office marketing, merchandising, customer service; self-checkout. Measured change in 2026: 74,000 retail jobs lost in the year to Feb 2026 (BRC). Entry-level retail assistant hiring +25% (DSIT/LinkedIn). Evidence quality: Good on job losses, but they are cost-driven. Automation is a response to labour costs, not the main cause.

  • Creative and media. Expected near-term task change: Copy, design, image and video production, localisation. Measured change in 2026: Creative industries filled jobs 2.464m in 2025, no significant change on 2024 (DCMS). Entry-level graphic designer hiring −28%. Evidence quality: Weak. Headline employment is flat, there are no 2026 income surveys, and only one entry-level signal.

What the measured changes have in common

They show up in hiring, not firing. The Bank of England’s July 2026 Monetary Policy Report says AI adoption “is gradually reducing demand for highly automatable jobs in some industries, with firms often slowing hiring or leaving vacancies unfilled”. A Bank Underground staff blog post (August 2026) finds vacancies down 15% in the most AI-exposed occupations over three years, against 10% for medium exposure and 6% for low exposure. Redundancies, meanwhile, were 3.6 per 1,000 employees in April–June 2026, “largely unchanged over the year”. If AI is changing jobs, it is doing so mainly by not refilling them.

They show up at the bottom of the ladder. DSIT and LinkedIn tracked 38 entry-level occupations in April 2026 and found 30 shrinking and 8 growing. The growing ones are retail assistants (+25%), sales development representatives (+17%) and business development representatives (+16%). The shrinking ones are the desk jobs: accountants, graphic designers, software engineers, product managers, data analysts. Their index of labour-market tightness puts “knowledge sectors” (tech, finance, professional services) at about 0.65–0.70, while “presence sectors” such as construction, hospitality and retail “remain near or above 1.0”. The government’s own caveat applies: the steepest declines are in areas “where AI has become most visibly capable, but further research is needed before conclusions can be drawn.”

They show up most clearly in sectors the official adoption survey cannot see. This is easy to miss. The ONS Business Insights and Conditions Survey is the main source on AI adoption. It excludes finance and insurance, public administration, and public provision of health and education. Those are three of the five sectors in this question. ONS’s July 2026 review finds most AI-using firms report “no change to their overall workforce headcount so far”, and just under 7% of medium-sized firms report a fall. That finding says nothing about banks, insurers or the NHS. For finance, the next systematic look is the joint Bank of England and FCA AI survey, which closed on 31 July 2026 and is due to report by the end of the year. The previous one, from November 2024, found 75% of financial firms already using AI.

Why the sectors diverge

Finance and professional services combine high wages, digitised inputs and existing quality-control habits, so a time saving is worth a lot and can be checked. They are also where executives are most willing to name AI as the reason. Standard Chartered’s chief executive talked of “replacing... lower-value human capital with the financial capital”. Large firms are also more likely to attribute cuts to AI generally: 60% of large employers that cut entry-level roles cite AI or automation, against 25% of small firms (Work Foundation, Aug 2026, reported).

IT is the clearest case of AI raising productivity and changing hiring together. Entry-level postings are down, but Adzuna’s average advertised IT salary rose 16.8% in the year to July 2026. That fits a sector paying more for fewer, more experienced people.

Customer contact has the clearest vacancy decline, but it also has the strongest non-AI explanation. British Gas cites a 20% fall in inbound calls as customers move online. Chatbots and digital channels are part of the same shift, so the argument over attribution is partly about definitions.

The public sector and the NHS have high exposure on paper, because so much of the work is correspondence, notes and casework. But headcount is set by budgets, restructuring and political choices, not tool adoption. Abolishing NHS England and cutting about 18,000 admin posts (reported) is a restructuring decision, not an AI one. Time saved by ambient voice tools is being counted as clinical capacity, not as posts removed.

Retail and creative industries show the danger of reading every job loss as AI. Retail lost 74,000 jobs, but 84% of retail finance leaders ranked labour costs among their top three concerns. Creative employment is flat overall, with the pressure concentrated in specific freelance and junior roles that official totals don’t separate out.

Bottom line

Expect the most task change, and the most measurable hiring change, in IT, customer contact, finance and the junior layers of professional services. In 2026 the measurable effect is a thinner intake rather than mass layoffs, and the evidence for that is moderate at best. In the public sector and the NHS, AI is changing tasks without changing headcount, which is set by other forces.

What to watch: the Bank of England and FCA survey results (end-2026), the ISE Student Recruitment Survey (14 October), the next ONS BICS wave with AI questions, and whether Bank staff update their vacancy analysis. Any of these could change where a sector sits in the list above.

Sources

From AI and Jobs: UK, October 2026