AI in UK business statistics

How many UK businesses use AI?

38.1% of UK businesses with 10 or more employees reported using at least one AI technology in September 2026, up from 34.9% in June. In September 2023 the same survey found 11.9%.

The figure is from the ONS’s Business Insights and Conditions Survey, which has asked the same question since 2023, so the rise is like for like. It counts a business as using AI if it uses any one of a list of AI technologies, and it does not tell light use from heavy.

Size changes the answer. 28.7% of businesses with 0 to 9 employees reported using AI in September 2026, against 50.9% of those with 250 or more. Across businesses of every size the figure was 29.7%, which is the one the ONS’s bulletin leads with.

Why do the headline figures disagree?

Because they are not the same measurement. The published figures run from 16% to 71%, and each one is right about the businesses it asked.

Four things move the number. Who was asked: a survey that leaves out the smallest firms reads higher, and so does one drawn from a membership. Who answered: the person in charge of technology cannot see the tools staff use on their own, and a finance director answers for the whole firm. How it is counted: one business, one vote gives a lower figure than weighting each business by how many people it employs. And when: the same panel went from 61% to 71% in under a year.

For the share of UK businesses using AI, the ONS figure is the one to quote: it has the largest sample, it is weighted to the business population, and it is repeated on the same question. Say which size of business it covers, because the ONS publishes one figure for businesses with 10 or more employees (38.1%) and another for businesses of every size (29.7%).

Which UK businesses use AI most?

Larger businesses, and businesses in information and communication, where 30.2% use one AI technology and a further 28.1% use two or more. Construction is lowest, at 12.4% and 0.6%.

London is ahead of the country as a whole. In DSIT’s 2025 survey, 20% of businesses in London were using AI, against 16% across the UK.

The same survey found 36% of large businesses using AI, 23% of mid-sized businesses and 14% of micro businesses, which it counts as 5 to 9 employees.

What do UK businesses use AI for?

Mostly to improve how the business already runs. Among businesses that use AI, 58.7% of those with 0 to 9 employees use it to improve business operations, and 63.8% of those with 250 or more.

Smaller businesses are more likely to point AI at customers and at new work. 31.2% of the smallest use it to provide or personalise products or services, against 18.1% of the largest, and 17.1% use it to explore a new market, against 7.7%.

By technology, large language models lead: 18% of businesses with 10 or more employees use them. Then come visual content creation (16%), data processing using machine learning (12%) and image processing using machine learning (6%). Robotics (2%) and other AI technologies (2%) are the least used.

DSIT’s survey of a year earlier found the same shape among businesses using AI: 85% used natural language processing and text generation, and 7% used agentic AI.

How deeply do businesses use AI?

Not very, so far. Businesses with 10 or more employees that use AI used an average of 1.59 AI technologies in June 2026, against 1.42 in September 2023.

Among businesses with 10 or more employees that use AI, 10% say they use it extensively. The smallest businesses are ahead on this measure: 17% of AI users with 0 to 9 employees, against 9% of those with 250 or more.

People are ahead of their employers. 55% of employed and self-employed adults in Great Britain say they use AI for work or education, against 34.9% of businesses with 10 or more employees that reported using it in June 2026. The two figures come from different surveys, which the ONS sets side by side, and the first includes use outside formal work.

In DSIT’s survey, businesses using AI said 30% of their staff use it on average. 53% of those businesses use AI constantly and a further 27% at least once a week.

Has AI changed jobs, productivity or revenue?

For most businesses, not so far. 89% of UK firms in the Bank of England’s Decision Maker Panel report no impact from AI on employment over the past three years, and 89% report no impact on productivity.

Where there is an effect on jobs, it follows what AI is used for. Among businesses that use AI to improve business operations, 63.1% report no change in headcount, 6.2% a reduction and 1.1% an increase. Among businesses using AI for other purposes only, the figures are 51.6%, 2.7% and 1.8%.

Businesses that use AI describe its effect more warmly than the panel does. In DSIT’s survey 75% of them said AI had improved the productivity of their workforce and 57% that it had given them new or improved processes. On revenue, 77% had seen no change since adopting AI and 12% an increase; 10% reported no impact at all so far. These are the businesses’ own reports, of businesses that already use AI, and the panel’s figure is for all firms and measures productivity as sales per employee.

In the BCC’s survey, 95% of the businesses using only generic AI tools reported no impact on headcount over the past 12 months.

What stops UK businesses adopting AI?

Most often, not seeing a use for it. In DSIT’s survey 71% of businesses said they had not identified a use for AI in their organisation, and 60% named limited AI skills, expertise or knowledge.

Cost comes well down the list, at 23%. In the same survey 80% of businesses neither used AI nor planned to, 5% planned to adopt it, and 51% did not see AI as relevant to their organisation.

A year on, the ONS found 41% of businesses with 10 or more employees reporting no barriers to adopting AI in the last three months.

How are businesses building AI skills?

By training the people they already have. 62% of businesses that name a lack of AI expertise as a barrier are training or retraining existing staff, against 26.4% of businesses that report no barriers.

Hiring for AI skills is rare in both groups, at 2.6% and 2.2%. Training has not reached most people yet: only 11% of businesses with 10 or more employees say more than half of their workforce has had AI-related training.

What do businesses plan to do with AI next?

Buy it ready-made, then build it into what they already run. Among businesses that use or plan to use AI, 65% plan to invest in off-the-shelf AI applications over the next 12 months and 59% plan to embed AI into their current applications and processes.

44% plan to augment human tasks with digital labour and 22% to build bespoke AI solutions. A business could choose more than one, so the figures do not add up to a whole.

How was this page made?

This page was researched and written with AI, and every figure on it was checked against its original source. Each source was opened on the date shown, and each figure found in it and copied exactly: from the data file under an ONS chart, from a report’s own figures and base lines, or from a working paper’s text. Nothing was taken from a press release, a news story or another list of statistics.

A figure is one published number. The count in this page’s title is the number of them the page holds, counted from the file the page is drawn from; a number shown twice is counted once. Where a chart adds two published figures together it says so, and the sum is not counted. Sample sizes and dates describe a source and are not counted.

Figures are shown as published, never rounded to suit a sentence. Surveys of different businesses, in different months, are never added or averaged.

Figures last checked: 21 October 2026

The sources

• Business insights and impact on the UK economy: 8 October 2026Office for National Statistics. Released 8 October 2026.Business Insights and Conditions Survey, Wave 165. UK businesses that answer a voluntary fortnightly survey. Of 38,592 businesses sampled for Wave 165, 9,933 responded, a response rate of 25.7%. The survey leaves out agriculture, oil and gas extraction, energy generation and supply, public administration and defence, public provision of education and health, and finance and insurance. Weighted by count to the business population. The figure for 10 or more employees leaves out the smallest businesses.The CSV file under the bulletin’s Figure 1.
• AI Adoption ResearchDepartment for Science, Innovation and Technology. Released 28 January 2026.Telephone survey by IFF Research and Technopolis Group. 3,500 UK private sector businesses with at least 5 employees. The interviewer asked for the person responsible for technology, so the report says it cannot see AI that staff use without the business knowing. Weighted to be representative by business size and sector.The figures and their base lines in the HTML report.
• Firm Data on AI (revised March 2026)Yotzov and others, NBER Working Paper 34836, on the Bank of England’s Decision Maker Panel. Released 19 March 2026.Decision Maker Panel, a monthly online survey run by the Bank of England. Senior executives of UK businesses with 10 or more employees: over 90% of the panel are chief financial officers, chief executives or other senior managers. The panel yields about 2,500 responses a month. Weighted by employment, so a large employer counts for more than a small one.The March 2026 revision of the working paper, section 4 and the notes under its figures. A revision of October 2026 adds a May to July 2026 wave, and this page does not yet use it.
• Artificial intelligence in UK businesses: 2023 to 2026Office for National Statistics. Released 20 July 2026.Business Insights and Conditions Survey, Wave 159. UK businesses that answer a voluntary fortnightly survey. Of 38,637 businesses sampled for Wave 159, 10,311 responded, a response rate of 26.7%. The survey leaves out agriculture, oil and gas extraction, energy, public administration and defence, public education and health, and finance and insurance. Weighted to the business population. The headline leaves out businesses with fewer than 10 employees.The CSV file under each chart, and the article’s own sentences. Wave 159’s sample and response are from the table in the ONS bulletin of 2 July 2026, because the article gives the sample as the number that responded.
• Powering Productivity: AI and the Future of UK WorkBritish Chambers of Commerce, with the University of Essex. Released 18 March 2026.BCC Business Outlook Survey, January 2026 wave. 668 businesses in the BCC’s membership network that answered the AI question: 561 with 1 to 250 workers, 68 sole traders and 39 with 250 or more. The authors say BCC members are slightly under-represented in London. Not weighted. The authors say the results describe the firms that answered, not UK SMEs as a whole.The report PDF, and the method in the working paper it cites (Bharier, Etheridge and Morais, ISER Working Paper 2026-01).

What we left out, and why

• The ONS’s headcount figures by size of business (its Figure 7). The chart’s title says they are businesses adopting AI; its notes do not say who was asked, and a third of the smallest firms answer “not applicable”. The base could not be told.
• The ONS’s figures for how businesses bring AI skills in, by size of business (its Figure 12). The article does not say whether every business was asked or only those using AI.
• The ONS’s 15% of businesses where more than half of employees use AI daily. The sentence follows one about businesses using AI and does not say whether it shares that base.
• The ONS’s barriers by size of business (lack of expertise, cost). They are in an interactive chart with no file to download, and the article gives them only as “around”.
• The BCC’s adoption figures for 2023, 2024 and 2025. The report charts them without a sample size or survey dates for those years.
• The BCC’s “95% of SMEs using AI report no impact on workforce size” and “86% say job roles are unchanged”. Both are from the press release. The report gives the 95% for users of generic AI tools only, which is the version kept here; the base of the 86% was not found.
• The share of UK and US workers using generative AI, from a Federal Reserve Bank of St. Louis report. Seen only as the ONS quotes it. The original was not opened.
• The Bank of England’s own summary of what its panel expects AI to do to jobs and productivity. Seen only in a search result. The Bank’s own page was not found, so the panel’s figures here are from the working paper, which was opened and read.

How to cite this page

SBW Labs. “AI in UK business statistics.” Figures last checked 21 October 2026. https://sbwlabs.com/research/ai-in-uk-business-statistics

For a single figure, cite the original publication it came from, linked under each chart.