Preparing your workspace
Jobbit
Guides22 min read

UK Small Business AI Statistics 2026: Who Uses AI, and for What

UK small business AI statistics for 2026 from ONS and government data: 29% of micro firms use AI against 51% of large firms, plus uses, barriers and jobs.

UK Small Business AI Statistics 2026: Who Uses AI, and for What
Read in:

About 3 in 10 UK businesses (30%) were using at least one artificial intelligence technology in September 2026, up from 9% in September 2023, according to the Office for National Statistics. Among businesses with fewer than 10 employees the share was 29%. Among businesses with 250 or more employees it was 51%.

Those three numbers are the short answer to the question this page exists to answer: how many UK small businesses use AI, and what do they use it for? The longer answer matters, because headlines on the subject range from "a third" to "more than half" depending on who asked the question and whom they asked. An owner who reads that 55% of small firms already use AI could reasonably conclude that they are behind. The official statistics say otherwise: about 7 in 10 of the smallest businesses have not reported using any AI technology at all.

This page collects the UK small business AI statistics that can be traced to a published table: the ONS Business Insights and Conditions Survey, the UK government's AI Adoption Survey, the Bank of England and, for comparison, Eurostat, the US Census Bureau and Statistics Canada. Membership and vendor surveys are included and labelled as such. Every figure names its source, its date and the population it covers. The page is reviewed each month as new data is published and was last updated on 15 November 2026.

Summary of findings

  1. First, adoption has roughly tripled in three years. The share of UK businesses using at least one AI technology rose from 9% in September 2023 to 30% in September 2026 in the ONS survey. Among businesses with 10 or more employees it rose from about 12% to 38%.
  2. Second, the smallest firms adopt least, and the gap in percentage points has widened. In September 2026, 29% of businesses with fewer than 10 employees used AI, against 51% of those with 250 or more. The gap between those two groups was 9.6 percentage points in September 2023 and 22.2 points in September 2026.
  3. Third, a second official survey agrees. The government's AI Adoption Survey 2026 found 32% of businesses with five or more employees using AI, double the 16% it recorded in 2025.
  4. Fourth, membership and vendor surveys report much higher figures, between 47% and 55%. They sample people who chose to answer, and they count any use of a general tool. They are not wrong, but they measure something different.
  5. Fifth, use is broad but shallow. Text generation (21%) and visual content creation (19%) lead among businesses with 10 or more employees. The main obstacle is skills, not price: 67% of adopters in the government survey reported limited AI skills, expertise or knowledge.
  6. Sixth, the effect on jobs so far is small. Among AI-adopting businesses with 10 or more employees, 6% told the ONS that adoption had reduced their headcount and fewer than 1% that it had increased it. The Bank of England's panel of firms reports close to zero effect on employment so far.

UK small business AI statistics at a glance

StatisticFigureWho it coversSource and date
Businesses using at least one AI technology30%All businesses in the surveyONS, September 2026
The same measure three years earlier9%All businesses in the surveyONS, September 2023
Micro businesses using AI29%0 to 9 employeesONS, September 2026
Large businesses using AI51%250 or more employeesONS, September 2026
Businesses with 10 or more employees using AI38%10 or more employeesONS, September 2026
Businesses with five or more employees using AI32%, up from 16% in 2025Private sector, 5 or more employeesUK government AI Adoption Survey, January to April 2026
Businesses planning to adopt AI in the next three months18%All businesses in the surveyONS, March 2026
Most used AI technologiesText generation 21%, visual content 19%10 or more employeesONS, September 2026
Adopters whose headcount fell because of AI6%Adopters with 10 or more employeesONS, September 2026
Adopters reporting an overall rise in productivity50%Adopters with 5 or more employeesUK government AI Adoption Survey 2026
Employees using AI for work or education55%Employees in Great BritainONS, May to June 2026
Private sector businesses in the UK5.6 million, of which 5.54 million have fewer than 50 employeesAll private sector businessesBusiness population estimates, start of 2026

The case: why the UK is a good laboratory for small business AI adoption

Most European countries measure business use of AI once a year. The UK measures it roughly every quarter. The ONS has asked the same question in its Business Insights and Conditions Survey since September 2023: is the business currently using at least one of a list of AI technologies? Because the survey runs every fortnight with rotating topics, there are now 13 readings across three years, each broken down by five employment size bands. That lets the UK series show how adoption moves within a year, and it publishes a figure for businesses with fewer than 10 employees. Eurostat, by contrast, excludes enterprises with fewer than 10 staff altogether.

The stakes are large because the UK is a small business economy. The government's business population estimates count 5,586,835 private sector businesses at the start of 2026. Of those, 5,539,260 (99.15%) have fewer than 50 employees and 4,115,175 (74%) have no employees at all. Only 8,355 have 250 or more. Any national statistic about "businesses" is therefore, by count, a statistic about very small firms.

The series has real limits, and they shape how the numbers should be read. The answers are self-reported by whoever completes the form, who may not know every tool staff use. The definition is broad: one person using a chatbot to draft emails counts the same as an automated production line. The survey covers most of the private sector but leaves out some industries, including agriculture, finance and the public sector, and it samples from the official business register, so many unregistered sole traders sit outside it. The ONS analysis published in July 2026 drew on 38,637 responding businesses, a response rate of 26.7%. Firms that are interested in technology may be more inclined to answer questions about it, although the ONS weights its results to reduce that risk.

A timeline of the evidence, 2023 to 2026

YearSourceSetting and designHeadline finding
2023ONS Management and Expectations SurveyOfficial survey of UK firms, reference year 20239% of firms had adopted AI
2023ONS Business Insights and Conditions SurveyFirst AI wave, September 20239% of all businesses used AI
2024ONS Business Insights and Conditions SurveySeptember 202414.7% of all businesses
2025UK government AI Adoption SurveyTelephone survey, 5 or more employees16% used AI
2025ONS Business Insights and Conditions SurveySeptember 202523.5% of all businesses
2025EurostatEU enterprises with 10 or more employees19.95% used AI
2026British Chambers of CommerceMember survey, 668 firms, unweighted54.3% used AI
2026UK government AI Adoption Survey3,503 interviews, January to April 202632% used AI
2026Statistics CanadaCanadian businesses, second quarter of 202619.2% used AI in the past 12 months
2026ONS analysis articleBusinesses with 10 or more employees, June 2026Use rose from around 12% to around 35%
2026ONS Business Insights and Conditions SurveySeptember 202630% of all businesses, 29% of micro businesses
2026US Census BureauBusiness Trends and Outlook Survey, October to November 202624.2% of firms used AI in the last two weeks
2026Bank of EnglandDecision Maker Panel, November 2026 reportClose to zero effect on employment so far

Finding 1: adoption has tripled since 2023, but the latest quarter was slow for the smallest firms

The ONS publishes the unrounded values behind its headline. The table below gives the share of businesses using at least one AI technology, by employment size band, at six points in the series.

DateAll businesses0 to 9 employees10 to 4950 to 99100 to 249250 or more
September 20239.4%9.1%11.3%14.5%15.0%18.7%
September 202414.7%14.4%16.7%23.2%25.6%30.4%
September 202523.5%23.1%25.8%33.4%35.1%39.4%
March 202625.9%25.1%30.3%39.3%42.1%44.6%
June 202628.9%28.2%33.1%42.7%45.1%48.6%
September 202629.7%28.7%36.1%45.8%50.4%50.9%

Three things stand out. The rise is steady, not explosive: 5.3 percentage points for all businesses in the first year, 8.8 in the second and 6.2 in the year to September 2026. Every size band has at least doubled. And the "all businesses" line sits almost on top of the micro business line, because businesses with fewer than 10 employees so heavily outnumber the rest.

The most recent quarter deserves attention. Between June and September 2026 the share rose by 3.0 points among businesses with 10 to 49 employees and by 5.3 points among those with 100 to 249, but by only 0.5 points among micro businesses. One quarter is not a trend, and a movement of half a point is within the noise any survey carries. If the pattern repeats in the December reading, it would suggest that the easy early adopters among the smallest firms have been reached and the remainder need more than access to a chatbot.

Intentions point the same way. In March 2026, 18% of businesses told the ONS they planned to adopt AI in the next three months. The realised increase over the following two quarters was 3.8 points. Stated plans run well ahead of what businesses then do, which is a reason to treat any forecast built on intentions with care.

Finding 2: the size gap is the central fact, and it has widened in percentage points

In September 2023, 9.1% of micro businesses and 18.7% of large businesses used AI, a gap of 9.6 percentage points. In September 2026 the figures were 28.7% and 50.9%, a gap of 22.2 points. Measured as a ratio the picture is different: large firms were about twice as likely as micro businesses to use AI in 2023 and a little under twice as likely in 2026. Both statements are true. The proportional gap has narrowed slightly while the absolute distance between small and large has more than doubled.

The government's AI Adoption Survey shows the same gradient with a different method. Among private sector businesses with five or more employees it found adoption of 27% for those with 5 to 9 employees, 34% for 10 to 49, 47% for 50 to 249 and 52% for 250 or more. Two independent official surveys, one online and one by telephone, land within a few points of each other in every band.

The gradient is international. Eurostat reports that 17% of small EU enterprises used AI in 2025, against 30.36% of medium and 55.03% of large ones. An OECD paper prepared for G7 ministers in December 2025 put the figures at 11.9% for firms with 10 to 49 employees and 40% for large firms. The size gap is not a British peculiarity. It appears wherever it is measured.

Finding 3: why one survey says 30% and another says 55%

A reader searching for the share of UK small businesses that use AI will find at least five different answers published in 2026.

SourceShare using AISampleType of survey
ONS, September 202630% (29% of micro businesses)Tens of thousands of businessesOfficial, weighted to the business population
UK government AI Adoption Survey, 202632%3,503 businesses with 5 or more employeesOfficial, weighted to the business population
Simply Business, September 202647%1,842 UK small business ownersCompany survey
British Chambers of Commerce, January 202654.3%668 firms, unweightedMembership survey
Federation of Small Businesses, November 202555%904 small business ownersMembership survey

The differences are not errors. They follow from three design choices. The first is who answers. The ONS and the government survey sample from business registers and databases and weight the results to the business population. The British Chambers of Commerce data, analysed by researchers at the University of Essex, comes from members who chose to respond, and the authors state plainly that the results are unweighted and describe the responding sample, not UK small businesses as a whole. The same caution applies to the Federation of Small Businesses figure and to the Simply Business survey of small business owners. Owners who join business organisations and answer technology surveys are likely to be more engaged than the average.

The second choice is what counts as use. The ONS asks about the business using specific technologies. Membership surveys tend to ask the owner whether they use AI tools, which captures a sole trader who asks ChatGPT to tidy a quote. The third is timing, which matters when the underlying figure moves by half a point a month.

The ONS itself shows how much the unit of measurement matters. In mid 2026, 55% of employees in Great Britain said they used AI for work or education, while about 35% of businesses with 10 or more employees reported using any AI technology. Staff are ahead of their employers. Much of the AI use in British workplaces is informal: individuals using general assistants on their own initiative, without the business having decided anything. For an owner, that is the practical lesson of this finding. The question is less whether AI is used in the business than whether anyone has decided how.

Finding 4: what small businesses use AI for, and what holds them back

Among businesses with 10 or more employees, the most used technologies in September 2026 were text generation with large language models (21%) and visual content creation (19%). The ONS analysis of June 2026 data ranked the rest: machine learning for data processing (12%), machine learning for image processing (6%) and robotics (2%). The description the ONS chose for the overall picture was that adoption has been "relatively shallow".

Purpose differs by size in an instructive way. Improving business operations is the leading purpose everywhere, cited by 58.7% of micro business adopters and 63.8% of large ones. But micro businesses that use AI are much more likely to point it at customers: 31.2% use it to provide or personalise products or services, against 18.1% of large adopters, and 24.4% use it to develop a new product or service, against 13.7%. One plausible reading is that a five-person firm has no back office to automate, so AI goes straight into what it sells.

The roles most affected, according to adopters with 10 or more employees, are administrative and clerical (38%), creative or design (24.1%), data analysis (21.9%), managerial (17.6%) and IT (14.4%). The government survey agrees on where AI lands first: administration (81% of those using or planning to use AI), marketing (71%) and IT (66%). It also found that 57% of adopters use AI constantly and a further 27% at least weekly.

By industry, information and communication leads at 60% in September 2026. Combining the categories the ONS published for June 2026 puts construction at about 13%, transport and storage at about 19% and wholesale and retail at about 21%, the three lowest. Construction is the industry of most tradespeople, which is worth remembering whenever a national average is quoted.

On barriers, the evidence is consistent across sources and somewhat surprising. Cost is not the main obstacle. In the ONS data, 41% of businesses with 10 or more employees reported no barrier at all, and cost was cited by between 7% and 14% depending on size. In the government survey, limited skills, expertise or knowledge affected 67% of adopters and 78% of those planning to adopt, and 65% of non-adopters said they lacked a clear business case. Eurostat finds the same across the EU: among enterprises that considered AI and did not use it, 70.89% cited a lack of relevant expertise. In the same survey the share of businesses calling AI "not relevant" fell from 51% in 2025 to 38% in 2026, so the remaining obstacle is less disbelief than not knowing where to begin.

Finding 5: jobs and productivity effects are small so far

The ONS asked adopters directly in September 2026. Among businesses with 10 or more employees that used AI, 6% said adoption had reduced their overall headcount and fewer than 1% said it had increased it. The Bank of England's November 2026 Monetary Policy Report reaches a similar conclusion from its Decision Maker Panel: firms report close to zero effect of AI on employment so far, and estimate that AI has raised their productivity by about 0.5% over the past three years. The Bank's regional agents add a nuance worth noting: AI is making firms less likely to hire or replace workers, without yet making them materially more likely to dismiss anyone.

The United States shows the same pattern. In the Census Bureau's survey for late October 2026, 87.3% of firms asked about the effect of AI on their headcount over six months reported no change, 6.6% an increase and 2.5% a decrease.

On productivity, half of adopters (50%) in the UK government survey reported an overall increase, which the report describes as generally modest. Of those able to estimate their spending, 47% had spent nothing on AI in 2025. A quarter (24%) said operating costs had fallen and 64% saw no immediate effect on costs. Low spending and modest gains belong together: most small firms are using free or low-cost general tools for individual tasks, not rebuilding how the work is done.

Controlled studies explain why the gains vary. In a study of customer support agents by Brynjolfsson, Li and Raymond, access to an AI assistant raised issues resolved per hour by 14% on average and by 34% for novice workers, with minimal effect for the most experienced. An experiment with Kenyan entrepreneurs by Otis and colleagues is closer to the small business case and more sobering. A GPT-4 business assistant had no detectable average effect on revenue or profit. Stronger entrepreneurs appeared to gain, while those who were already struggling did nearly 10% worse, apparently because they sought advice on harder problems. Firm-level research from the Bank for International Settlements finds that AI adoption raises labour productivity by 4%, with the gains concentrated in medium and large firms. The honest summary is that AI helps most where a task is well defined and the user can judge the output, and that small firms have not yet captured the gains larger ones report.

Mechanisms: why the pattern looks like this

Low entry cost produces broad, shallow adoption. A general assistant costs nothing or a few pounds a month and needs no installation. That explains both the speed of the rise and the ONS description of adoption so far as relatively shallow.

Fixed costs favour size. Connecting AI to stock, bookings or accounts, cleaning the data it needs and checking its output are costs that do not shrink with the size of the firm. A business with 300 staff spreads them across many tasks. A business with three cannot, which fits the widening gap in percentage points.

Knowing what to ask for is the scarce skill. Every source that asks about barriers finds expertise and use cases ahead of price. In 2023 the top barrier in the ONS management practices study was difficulty identifying activities or business use cases, cited by 39%, with cost at 21%. Three years later the wording has changed and the finding has not.

Use is led by employees, not by decisions. With 55% of employees using AI against about 35% of businesses reporting it, much adoption happens before the owner has set any rule about data, accuracy or disclosure.

A sceptic should hold some alternative explanations alongside these. Part of the measured rise may be recognition, not new behaviour: software that businesses already used has added AI features, and respondents have become more likely to notice and report them. Question wording has shifted in some surveys, most visibly in the US series, which is why this page avoids comparing its 2023 and 2026 values directly. And response rates near a quarter leave room for the firms that answer to differ from those that do not.

How the industry responded

Government has treated the size gap as a policy problem. The SME Digital Adoption Taskforce, reporting in July 2025, set the ambition that UK small and medium-sized businesses should be the most digitally capable and AI confident in the G7 by 2035, and the government has now run its AI Adoption Survey in two consecutive years to track progress.

Technology companies have taken a different route: putting AI inside products small firms already pay for. Microsoft has built Copilot into its office software, Google has done the same with Gemini, and accounting platforms such as Xero, Sage and Intuit QuickBooks have added assistants of their own. Their surveys report high adoption. Intuit says 77% of US small and mid-sized businesses use AI regularly, and the US Chamber of Commerce puts the figure at 66%. These are vendor and trade body surveys and should be read with the cautions set out in Finding 3. A Xero survey of US owners offers a useful counterweight: 53% said their business would not be affected if AI tools disappeared tomorrow.

A third response is the AI agent: software that does not just answer a question but carries out a piece of work, such as building a website, chasing invoices or compiling research, and shows what it did. Jobbit belongs to this group. Its design fits the evidence above in two specific ways. The agent works on digital tasks whose output the owner can inspect and correct, which is where the studies find reliable gains, and tasks that need a person can be passed to vetted human professionals on pro.jobbit.uk. Jobbit has not published trial data of its own, so that is a statement about design, not about measured results. Readers who want to see what agent work looks like in practice can start with AI use cases for small business, the first week with an AI agent and our earlier case study on whether AI agents replace freelancers.

What a business should do with this evidence

Do not measure yourself against the highest headline. The official figure for businesses with fewer than 10 employees is 29%. If you have not started, you are in the majority, and the firms that have started are mostly using a chatbot for drafting.

Start where the evidence says the gains are: administrative and clerical work with a clear right answer. Quotes, invoice reminders, replies to routine enquiries, summaries of long documents. Our list of AI automation ideas for small businesses gives 50 concrete examples.

Decide before your staff decide for you. More than half of employees say they already use AI for work or education. A one-page policy that says which tools are allowed, what must never be pasted into them and who checks the output costs an hour to write, and our AI policy template can be copied. General assistants such as ChatGPT, Claude, Gemini, Microsoft Copilot and Meta's Muse all need the same rules.

Keep a person on anything involving money, legal commitments or safety, and on any advice you cannot check yourself. The Kenyan experiment is the warning: owners who were already struggling did worse with an AI adviser, not better.

Measure one task for one month. Time it before and after. If the saving is real, extend it. If it is not, stop, and try a different task.

Working rule: give AI one routine task whose result you can check in a minute, measure it for a month, and expand only what the numbers support.

Limitations and open questions

Every adoption figure on this page is self-reported, and "using at least one AI technology" says nothing about how often, how well or to what effect. The official UK surveys do not fully cover businesses with no employees, which are 74% of the business population, so the statistics describe small employers better than sole traders. The industry figures for June 2026 combine two published categories and are rounded here for that reason. International comparisons use different populations, reference periods and question wordings and should be read as context, not as a league table. The Federation of Small Businesses figures are taken from press coverage of its report.

Three questions remain open. Whether the slowdown among micro businesses in the September 2026 reading persists. Whether productivity gains reported by half of adopters appear in output statistics. And whether the hiring effect the Bank of England's agents describe, fewer replacements without more dismissals, grows into something visible in employment data. The next ONS readings will help with the first. The others will take years.

Conclusion

Who uses AI among UK small businesses, and for what? On the best available evidence, about 29% of businesses with fewer than 10 employees and 36% of those with 10 to 49 used at least one AI technology in September 2026, mainly to generate text and images and to speed up administrative work. Adoption has tripled since 2023, larger firms remain far ahead, and the measured effects on jobs and productivity are modest. Surveys reporting that more than half of small businesses use AI are measuring engaged owners using general tools, which is a real but different thing.

For a small business the practical conclusion is reassuring. There is no race already lost. The firms gaining most are not those that adopted first but those that chose a task they understand, checked the output and kept score.

Frequently asked questions

What percentage of UK small businesses use AI?

According to the Office for National Statistics, 29% of UK businesses with fewer than 10 employees and 36% of those with 10 to 49 employees used at least one AI technology in September 2026. Across businesses of all sizes the figure was 30%. A separate government survey of businesses with five or more employees found 32% in early 2026.

How many UK businesses use AI in 2026?

About 3 in 10. The ONS reported that 30% of all businesses used at least one AI technology in September 2026, up from 9% in September 2023. Among businesses with 250 or more employees the share was 51%.

Why do some surveys say more than half of small businesses use AI?

Surveys by membership bodies and companies, such as the British Chambers of Commerce at 54% and the Federation of Small Businesses at 55%, question people who chose to respond and count any use of a general AI tool by the owner. Official surveys use random samples and ask about the business using specific technologies. Both are valid, but only the official figures describe the whole business population.

What do small businesses use AI for?

Mostly text generation and image creation, used for administrative work, marketing and routine customer communication. In the ONS data, 58.7% of micro business adopters use AI to improve business operations and 31.2% to provide or personalise products or services. Administrative and clerical roles are the most affected.

Is AI replacing jobs in UK small businesses?

Not on any large scale so far. In September 2026, 6% of AI-adopting businesses with 10 or more employees told the ONS that adoption had reduced their headcount, and fewer than 1% said it had increased it. The Bank of England reports close to zero overall effect on employment so far, though firms are somewhat less likely to hire or replace workers.

Which UK industries use AI the most?

Information and communication leads, with 60% of businesses using AI in September 2026. Education and professional, scientific and technical services follow. Construction, transport and storage, and wholesale and retail have the lowest rates, at roughly 13% to 21% in mid 2026.

Related guides