by Claude Opus 5.5

What does UK job-posting evidence up to October 2026 suggest about which skills and roles are rising or falling in demand, how (if at all) is AI referenced, and what has shifted since January?

UK job postings fell further in 2026, and demand narrowed. Ads for AI, data, technology, engineering, trades and teaching held up or grew. Hospitality, retail, logistics and junior information-processing roles (accountants, designers, software engineers, analysts) fell. AI is now mentioned in a record 9.4% of UK postings, and the mentions have spread well beyond tech jobs. Most of these ads ask people to use AI rather than build it.

The backdrop: fewer ads than the official data suggest

Indeed’s mid-year update (3 August 2026) puts UK postings 32% below their February 2020 baseline and 11% lower than at the start of 2026. The euro area and the US remain at or above their pre-pandemic levels. Adzuna counted 791,490 live vacancies in July 2026, down 9.6% on the year. Indeed’s Jack Kennedy calls the result “a two-speed labour market, with weak overall hiring but rapidly growing demand for a new set of skills.”

Online postings have fallen faster than the ONS vacancy survey, which shows a fall of about 4% between October–December 2025 (734,000) and June–August 2026 (702,000) (computed). Advertising is the cheapest thing for an employer to cut, and with many applicants per role, some employers fill posts without advertising widely. Treat posting data as a measure of demand for new hires at the margin, not a count of jobs.

Where demand is rising

  • AI specialists. PwC’s 2026 AI Jobs Barometer (15 June 2026) found UK postings for specialist AI roles rose 61% in 2025, from 112,000 to 180,000, taking them to 2.2% of all postings. The advertised wage premium for AI skills was 34.2%, up from 11% in 2024. It ranges from 12% in government to 64% in consumer markets. LinkedIn (June 2026) counts 95,000 AI roles created in the UK since 2023.

  • Technology, data and engineering more broadly. Indeed reports that tech, engineering and healthcare are bucking the downward trend. IT systems roles show the fastest posted wage growth, at 7.2%, against 3.9% across all postings, the lowest overall rate since February 2022. Adzuna’s average advertised IT salary was £67,771 in July, up 16.8% on the year.

  • The physical economy. In Adzuna’s July data, teaching vacancies rose 11.9% on the year, trade and construction 8.3% and manufacturing 2.2%. The KPMG/REC Report on Jobs (September 2026) found permanent staff demand rising in engineering and in accounting and finance, and temporary demand rising in blue-collar and IT roles.

  • Customer-facing entry roles. A DSIT and LinkedIn snapshot of entry-level hiring (June 2026) found growth in retail assistants (+25%), sales development representatives (+17%) and business development representatives (+16%) while most other entry roles shrank.

Where demand is falling

  • Consumer-facing and logistics roles. Adzuna recorded hospitality and catering down 26.8%, logistics and warehouse down 28.1%, and healthcare and nursing down 30.6% on the year in July. KPMG/REC found demand weakest in retail and in hotels and catering. These sectors are the most exposed to the 2025 rise in employer National Insurance and to National Living Wage increases. That is a cost story, not an AI one. The healthcare figures are contested: Indeed has healthcare among its more resilient categories, a reminder that job boards cover different parts of the market.

  • Junior information-processing roles. The DSIT/LinkedIn snapshot found entry-level hiring falling in 30 of 38 tracked occupations, measured in April 2026 against a year earlier. The steepest falls were:

  • Accountant. Change in entry-level hiring: −29%.

  • Graphic designer. Change in entry-level hiring: −28%.

  • Software engineer. Change in entry-level hiring: −27%.

  • Product manager. Change in entry-level hiring: −24%.

  • Data analyst. Change in entry-level hiring: −15%.

  • Legal assistant. Change in entry-level hiring: −14%.

  • Administrative and customer-service work. A Bank of England staff blog post (Bank Underground, 6 August 2026) found vacancies in the most AI-exposed occupations down 15% over three years, against 10% for medium-exposure and 6% for low-exposure occupations. Customer-service and administrative vacancies fell by more than 20%.

  • Graduate roles. Indeed’s graduate postings were about 7% lower than a year earlier as of 10 July, the lowest for the time of year since 2020. Adzuna’s much steeper −45.6% is disputed (see 2.3).

How AI appears in the ads

Mentions are spreading. Indeed found 9.4% of UK postings mentioned AI or related tools at the end of June 2026, a record and well above the share Indeed reported in late 2025. It reaches 48.8% in data and analytics, with notable shares in software development, scientific research and marketing. In beauty, cleaning and driving it is below 1%. Kennedy’s summary: “AI literacy is no longer confined to specialist technology roles – it is spreading into marketing, media, finance, scientific research and beyond.”

Most AI demand is for users, not builders. PwC found UK “AI user” postings grew 65.8% (about 66,000 roles), against 21.6% for developer roles (about 2,600). The typical new AI requirement is a marketer, analyst or finance professional expected to work with AI tools, not a machine-learning engineer.

Entry roles are being “seniorised”, at least in the US. PwC’s analysis of US job ads finds the most AI-exposed entry-level roles are now “seven times more likely to require traditionally senior-level skills”. These US roles have grown 35% since 2019, while other US entry-level roles fell 10%. In that data, the entry-level job is not disappearing so much as being redefined upwards: fewer pure drafting jobs, more jobs asking juniors to review, check and exercise judgement from day one.

Jobseekers are following the demand. Indeed says searches for AI-related roles are up roughly sevenfold since ChatGPT launched.

One caution on mentions. A keyword in an ad is a signal of what employers want to be seen asking for. It is not proof that AI is doing work that people used to do. Some of the rise is employer branding, and some reflects AI tools becoming standard software, much as “Excel” once did.

What has shifted since January

The January edition described postings that were “soft and selective” but stabilising. Since then:

  1. Overall postings fell further, not sideways. Indeed’s postings dropped another 11% between January and July, leaving them 32% below the pre-pandemic baseline.

  2. Year-on-year falls in junior ads look milder, mainly because of the comparison. Adzuna’s entry-level vacancies were down 8.1% on the year in July 2026, and Indeed’s graduate postings 7%. Both are now being compared with summer 2025, which was already weak, so the smaller declines do not mean recovery.

  3. AI mentions reached a record and the advertised AI wage premium roughly tripled in PwC’s measure.

  4. Occupation-level evidence arrived. The DSIT/LinkedIn snapshot and the Bank staff exposure analysis allow, for the first time, a comparison between AI-exposed and less exposed roles. Both show steeper falls where AI is most capable.

  5. A credible counter-reading emerged. Bloomberg Economics (June 2026) analysed online vacancies for 400 job types and found that AI-exposed vacancies were already falling before ChatGPT and have risen since summer 2024. Its conclusion: “the results push against the idea of rapid, large-scale job displacement.”

Bottom line

The posting data show demand being reshaped, not collapsing: fewer junior information-processing roles, more AI-using and technical roles, and resilience in trades and teaching. The Bank staff and Bloomberg readings differ mainly because they use different exposure definitions and time windows. Both may be right: exposed roles fell further over three years, but not at an accelerating rate since 2024. Which pattern holds will be clearest when overall hiring recovers. If ads for exposed junior roles stay down while everything else picks up, the AI explanation will be much harder to dismiss.

Sources

From AI and Jobs: UK, October 2026