Start here Free tools
A carpenter working on a laptop in his workshop, surrounded by tools and timber

How many businesses actually use AI? What the 2026 data says

The short answer. About three in ten UK businesses used at least one kind of AI in September 2026, rising to 38% of those with ten or more employees, according to the Office for National Statistics. In the US, the Census Bureau puts it at 23.8% of employer businesses; in the EU, 19.95% of firms with ten or more employees used AI in 2025. Bigger firms use it more everywhere, and the main brake on smaller ones is not cost but not knowing what to use it for.

Every few weeks a new survey says most small businesses now use AI. The official statistics tell a slower and more useful story, and the gap between the two is worth understanding before you decide anything.

What do the official figures say?

Three national statistics bodies now ask businesses about AI on a regular basis. They do not ask exactly the same question, or ask the same kinds of business, so the numbers are not a league table. Read side by side, though, they agree on the direction and roughly on the scale.

WhereWho was askedLatestEarlierSource
UKAll businesses using at least one AI technology30% (Sep 2026)9% (Sep 2023)ONS, Business Insights and Conditions Survey, wave 165
UKBusinesses with 10 or more employees38% (Sep 2026)35% (Jun 2026)ONS, same survey
USEmployer businesses using AI in any business function in the last two weeks23.8% (latest fortnight)17.3% (late 2025, when the question changed)US Census Bureau, Business Trends and Outlook Survey
EUEnterprises with 10 or more employees19.95% (2025)13.48% (2024)Eurostat

Two things in that table are easy to miss. The EU figure leaves out every business with fewer than ten employees, which is most businesses. And the US figure cannot be compared with anything before November 2025, because the Census Bureau changed the wording of its question then and started a new series. Anyone claiming American AI use has doubled in a year is comparing two different questions.

Why do surveys disagree so much?

Because they count different businesses in different ways. A survey of a membership organisation reaches firms that are already engaged, connected and more likely to be trying new things. The British Chambers of Commerce found that 54.3% of the 668 member firms answering its January 2026 survey used AI. The researchers who analysed it, at the University of Essex, are clear that those results are unweighted and describe the firms that responded, not UK businesses as a whole.

The same is true in the US. The Federal Reserve’s Small Business Credit Survey, fielded in late 2025, found 46% of small employer firms said they or their employees currently use AI. The Fed itself notes the survey is not a random sample. Weighted national statistics, built to represent every business, come in lower: 30% in the UK, 23.8% in the US.

Definitions matter as much as samples. “Using AI” can mean one person drafting the odd email with a chatbot, or a business that has rebuilt a process around it. Most surveys count both the same way. So when you see a single headline number for “small businesses”, it is worth asking who was asked, how small “small” was, and what counted as use.

Do smaller businesses use AI less?

Yes, consistently, in every source that splits the figures by size. In the UK in September 2026, 29% of businesses with fewer than ten employees used AI, against around half of the largest businesses. In the EU, 17% of small enterprises (10 to 49 employees) used AI in 2025, against 55.03% of large ones. The UK government’s own adoption research, carried out in early 2025, found 14% of micro businesses using AI against 36% of large ones.

That gap is not a sign that small businesses are behind. Large firms have more repetitive work at scale, more people to try things, and more data in one place. A business of eight people has less of all three, which changes where AI pays.

What is holding businesses back?

Mostly not money. In the government’s AI Adoption Research, which surveyed 3,500 UK private-sector businesses with at least five employees between February and May 2025, the top barriers were a lack of identified need for AI (71%) and limited AI skills, expertise and knowledge (60%). Cost was cited by 23%.

Eurostat asks a narrower question: of the firms that had considered using AI but were not using it, what stopped them. Lack of expertise came top there too. Because the two surveys ask different groups, the percentages should not be compared, but the message is the same. The hardest part is not buying a tool. It is knowing which job it should do.

Is AI replacing jobs in smaller firms?

Not much, so far, on the official evidence. Of UK businesses with ten or more employees that were using AI in September 2026, 6% said it had reduced their headcount and fewer than 1% said it had increased it. The ONS describes both figures as broadly stable over the past year. Administrative and clerical roles were the ones businesses most often said had been affected, which usually means the work inside the role has changed rather than the role disappearing.

What should a small business take from this?

First, if you are not using AI yet, you are in the majority, not left behind. Around seven in ten UK businesses did not report using it in September 2026.

Second, the businesses that get value tend to start from a specific, repetitive task rather than from a tool. The barrier the data points to most often, not knowing what to use it for, is exactly the one you can fix on paper before spending anything. Pick one job that happens often and the same way, check the information it needs is somewhere your team can reach, give it one owner, and decide what success looks like as a number.

Our free AI readiness check walks through those questions in a few minutes, and the glossary explains AI readiness and data readiness in plain English.

Sources

Related reading

Back to Insights