AI adoption statistics for small and mid-sized businesses in 2026

Ask five sources how many businesses use AI and you will get five answers, from roughly one in five to nearly nine in ten. Both ends are defensible. The US Census Bureau reported that overall AI use hovered between 17% and 20% of US businesses between December 2025 and May 2026, while McKinsey found that 88% of surveyed organisations reported regular AI use in at least one business function in 2025. The numbers differ because they measure different things in different populations.

This guide puts the main published AI adoption statistics side by side, explains why they disagree, and gives a realistic range you can quote with confidence. It covers the United States, the United Kingdom, the European Union and global survey data, and breaks the figures down by company size and sector wherever the sources allow. None of the figures below are Klevere data. Each one comes from the publisher named next to it.

Quick answer

Official statistics put AI use at roughly 20% of businesses in the US and EU and around 35% in the UK, among firms of all but the smallest sizes. Large firms are far ahead: 37% of US firms with 250 or more employees and 49% of UK firms of that size. Surveys of larger organisations report close to 90%, because they sample differently.

The headline numbers at a glance

The list below shows the most widely cited sources. Read the scope and year before you quote any of them, because a figure without its denominator is close to meaningless.

United States: about one in five businesses

The most reliable US number comes from the Census Bureau's Business Trends and Outlook Survey, a biweekly, nationally representative survey. In its May 2026 analysis, the Bureau reported that the national rate was 19.8% as of 3 May 2026, and that between 20% and 23% of businesses expected to be using AI within six months. That points to steady growth rather than a sudden jump.

A Federal Reserve note by Jeffrey S. Allen reached a similar figure, finding that about 18 percent of firms had adopted AI as of year-end 2025. It also recommends the Census series as the best source for the share of US businesses that have adopted AI.

Size matters a great deal. The Census Bureau found that 37% of firms with at least 250 employees reported using AI in their operations, compared with 32% of firms with 100 to 249 employees. It also found that AI use rose among firms with 20 or more employees but did not change significantly among firms with fewer than 20, and that less than 20% of firms with four or fewer employees used it.

Sector differences are just as wide. As of May 2026, information businesses stood at 39.7% and finance and insurance at 33.9%, while retail trade sat at around 14%. Anyone writing about a specific industry should use the sector figure, not the national average.

One caveat applies to every US trend line. The Census Bureau changed its question in November 2025, from AI use in producing goods or services to AI use in any business function. A series that appears to jump or flatten around that date partly reflects the wording, not a change in behaviour.

United Kingdom: closer to one in three

The Office for National Statistics publishes the most detailed UK picture through the Business Insights and Conditions Survey. Its July 2026 release found that self-reported AI use in businesses with 10 or more employees rose from around 12% to around 35% between late 2023 and June 2026, with larger firms more likely to have adopted AI. That is just under a threefold increase in under three years.

The size gradient is clear. The ONS reports 28% of businesses with 0 to 9 employees using at least one AI technology, compared with 49% of those with 250 or more employees. Sector spread is wider still: almost three-fifths of information and communication businesses used AI, while construction was the lowest at 13%.

The ONS is candid about what the headline hides. It states that the measure treats all reported uses as equivalent and does not distinguish light use from embedded, production-level adoption. The depth data supports that caution. Average AI technologies per adopting business rose only from around 1.4 to around 1.6, and 10% of businesses with 10 or more employees reported extensive use. Large language models were the most common technology at 18%.

The survey behind this had 38,637 responding businesses in its latest wave, a 26.7% response rate, and excludes sectors such as finance and insurance. The ONS also labels these figures official statistics in development. Another useful contrast: 55% of employees reported using AI for work or education, compared with about 35% of businesses. People adopt faster than organisations.

European Union: 20% and rising fast from a low base

Eurostat's annual survey of enterprises with 10 or more employees found that 20.0% used AI technologies in 2025, a rise of 6.5 percentage points from 13.5% in 2024. The 2023 figure was 8.1%. The rate has therefore moved from roughly one in twelve to one in five in two years.

Country differences are large. In 2025 Denmark led at 42.0%, followed by Finland at 37.8% and Sweden at 35.0%, while Romania recorded 5.2%, Poland 8.4% and Bulgaria 8.5%. Almost every member state improved on the previous year. The most common uses were analysing written language at 11.8% and generating images, video or audio at 9.5%.

If you operate across several European markets, the EU average hides a spread of more than eight to one between the highest and lowest countries. Planning a rollout, a sales message or a compliance programme around the average would be a mistake. Our guide to EU AI Act obligations for SMBs covers the regulatory side of that variation.

Global and worker surveys: where the 88% comes from

McKinsey's global survey tells a different story because it asks a different population. Its 2025 edition found that 88 percent reported regular AI use in at least one business function, compared with 78 percent a year earlier. Respondents were 1,993 participants across 105 nations, and they skew towards people who know their organisation's AI activity well. That is not the same as a random sample of every business.

The depth gap shows up in the same report. Only about one-third said their companies had begun to scale their AI programmes, and 62 percent said their organisations were at least experimenting with AI agents, with 23 percent scaling an agentic system somewhere. The 2026 edition reports that 44 percent now say AI is scaling across their enterprise, up from 38 percent a year ago, and notes that agent adoption among smaller organisations stayed flat at 22 percent.

Worker-level data is higher again. The Federal Reserve note reports that work-related generative AI use stood at about 41 percent of the workforce in November 2025, and the ONS found 55% of UK employees using AI for work or education. Individuals often use tools their employer has not formally adopted, so employee numbers will always exceed business numbers.

Why the numbers disagree

Four factors explain nearly all of the gap between a 20% headline and a 90% headline.

The realistic range to quote

Putting the evidence together, a defensible summary for a business audience looks like this. These ranges are our reading of the sources above, not new data.

A sentence you can safely use is this: about one in five businesses use AI in the US and EU, about one in three in the UK, with large firms roughly twice as likely as small ones, and with most use still light.

What the statistics mean for a small or mid-sized business

Three practical conclusions follow from the data. First, being behind the average is normal. If you run a business with fewer than 20 employees, you are in the group where official figures show the least change over the last six months. Second, the competitive gap is in depth, not access. Nearly every firm can use a chatbot, but only a minority have embedded AI in a workflow that runs without someone prompting it. Third, sector matters more than country. A construction firm in the UK sits at 13%, while an information business in the US sits near 40%.

For a practical starting point, our AI for small business guide walks through where first projects usually pay back, and is my business ready for custom AI gives a short readiness check before you commit time or budget. Once a project is live, measure it properly: our piece on AI ROI explains how.

How Klevere approaches adoption

Klevere builds custom AI agents for small and mid-sized businesses, and the adoption data shapes how we work. Having deployed 500+ AI agents across 50+ projects and 12 industries, we see the same pattern the statistics describe: access to tools is rarely the constraint. The constraint is a clear, specific process that someone owns, with data the agent can reach and a way to check its output.

So we start with the process, not the model. We map where time is lost, choose one workflow with a measurable outcome, and connect the agent to the systems the team already uses. Human review stays in place for anything with legal, financial or reputational consequences, and we agree up front how success will be measured. Our AI audit is the structured way to do that first step, and it is designed to tell you honestly when a process is not worth automating.

We would also push back on the idea that you must match the headline numbers. Adopting AI because 88% of a survey sample says they have is a poor reason. Adopting it because a specific task costs you ten hours a week is a good one.

How we compiled this

We selected sources that publish a stated method: national statistics offices and central bank research (the US Census Bureau, the UK Office for National Statistics, Eurostat and the Federal Reserve), plus one large global survey from McKinsey with its sample size and fieldwork dates stated. We opened each page and took figures only where they appeared on that page. We did not average, adjust or combine figures from different surveys, and we flag differences in scope beside each number. Data was checked in November 2026 and will change as new releases appear, so check the publisher's page for the latest.

Frequently asked questions

What percentage of businesses use AI in 2026?

It depends on the source and the country. Official statistics show about 20% of US businesses (Census Bureau), 20% of EU enterprises with 10 or more employees in 2025 (Eurostat) and around 35% of UK businesses with 10 or more employees (ONS). Executive surveys report far higher figures.

How many small businesses use AI?

Fewer than larger firms. The Census Bureau found that less than 20% of US firms with four or fewer employees used AI, and the ONS found 28% of UK businesses with 0 to 9 employees using at least one AI technology. Very small firms also showed little recent change in the US data.

Why do AI adoption statistics vary so much?

They measure different populations, definitions and time windows. A survey counting every firm equally is dominated by tiny businesses, while an employment-weighted or executive survey reflects large companies. Broad definitions that include free chatbots give higher rates than definitions limited to production use. Always check scope before comparing numbers.

Which industries use AI the most?

Information and communication leads in both countries with official data. The Census Bureau recorded 39.7% in information and 33.9% in finance and insurance, and the ONS found almost three-fifths of UK information and communication firms using AI. Construction (13% in the UK) and retail (around 14% in the US) sit near the bottom.

Is AI adoption growing?

Yes, but at different speeds. UK use rose from around 12% to around 35% between late 2023 and June 2026, and EU use rose from 13.5% to 20.0% in the year to 2025. US growth has been slower recently, with use holding between 17% and 20% from December 2025 to May 2026.

Does using AI mean a business has scaled it?

No. McKinsey found that only about one-third of 2025 respondents had begun to scale AI, and 44 percent reported enterprise-wide scaling in 2026. The ONS likewise notes that its headline does not separate light use from embedded adoption. Most businesses are still in early or partial use.

If you want to know where your own business sits against these numbers, and which one or two workflows are worth automating first, book a free AI audit and we will walk through it with you.