UK firms hire more AI workers to manage bots

UK employers hired more AI workers in 2025 as companies looked for staff who can apply AI tools inside finance, marketing, operations, software, and public services, according to PwC.

PwC said in its AI Jobs Barometer that UK hiring for AI specialists rose 61%, from 112,000 roles in 2024 to 180,000 roles in 2025. The wider UK jobs market moved in the other direction, with overall vacancies down 6.6%.

The split gives compliance, risk, and technology leaders a practical signal. Companies need fewer pure model builders than the hype suggests. They need people who understand a business process, know where AI can help, and can check the output before a customer, regulator, auditor, or manager relies on it.

PwC analysts said employers added almost 66,000 AI user roles during the year. Employers added 2,600 AI developer roles. That gap matters for workforce planning because it puts AI governance inside ordinary business functions, rather than inside a narrow engineering group.

A bank that hires an AI-enabled fraud analyst still needs that analyst to understand fraud typologies, false positives, customer harm, and escalation rules. A law firm that hires AI-capable associates still needs lawyers to check privilege, citations, confidentiality, and client instructions. A manufacturer that gives planners AI forecasting tools still needs staff who can challenge bad assumptions before the company buys parts or misses orders.

PwC described a two-track labor market. Roles where workers use AI to remove repetitive tasks grew 39% since 2018, PwC said. Jobs where AI lowers the skill barrier grew 17% over the same period.

Workers with AI skills now earn an average wage premium of 34.2%, up from 11% a year earlier, PwC said. Consumer market companies offered the largest premium at 64%. Government and public sector employers offered a 12% premium.

The numbers give employers a pay and retention problem. A company that trains staff on AI tools may create a poaching risk. A company that refuses to train staff may create a control risk, because employees will still test public AI tools without clear approval, logging, or data rules.

Compliance teams should treat the hiring shift as an operating model issue. AI users need role-based training, approved tools, data handling rules, escalation routes, and audit trails. Managers need evidence that staff reviewed AI output before they used it in customer communications, financial analysis, legal work, hiring, procurement, or security decisions.

The UK figures also explain why office workers spend time checking AI output. PwC's numbers suggest companies have moved from pilots into daily use, but many employers still lack review standards. Staff can gain speed from AI drafts, summaries, and code suggestions, then lose time when they must repair wrong facts, weak reasoning, missing context, or confidential data exposure.

For developers and IT leaders, the hiring pattern changes the build-versus-buy question. A company may need a smaller group of AI engineers to configure models, connect systems, and monitor performance. It may need a much larger group of trained users who can write prompts, test outputs, spot failure modes, and document decisions.

That structure also changes accountability. Legal, security, HR, finance, and operations teams should own the risks inside their workflows. Central AI teams can set standards and provide platforms, but the business owner needs to decide whether a model output meets the standard for use.

Employers should review job descriptions before they chase the wage premium. A vague demand for AI skills will attract candidates who can name tools. A useful job description should name the workflow, the data types, the review duty, and the decision the worker will influence.

For workers, the premium points to a clear path. Domain expertise still matters. AI skills add value when a person can apply the tool to a specific job and catch mistakes before they reach production, customers, or regulators.

PwC benefits from the AI consulting market, so employers should read its conclusions with that context. The labor trend still fits what companies now face: AI projects need human reviewers, process owners, risk controls, and staff who can turn software output into work the business can defend.