Is everyone using AI except you? Not yet, but that’s your last window

How many businesses actually use AI?

Around one in five in the US, Canada and the EU, by their official statistics. The US Census Bureau found business AI use hovered between 17% and 20% from December 2025 to May 2026¹. Statistics Canada put it at 19.2% in the second quarter of 2026². Across the EU, 20.0% of enterprises with 10 or more employees used AI in 2025³.

Elsewhere, the official figures count different groups. In the UK, 27% of businesses that handle digital data use AI-based technologies⁴. In Australia, 43% of small and medium businesses report “some level” of AI adoption⁵. In Hong Kong, the share of firms using AI rose from 7% in 2021 to 20% in 2025⁶. South Korea’s government survey found 28% of companies using AI, though back in 2023⁷. And in China, the Ministry of Industry and Information Technology says AI adoption among large industrial enterprises has passed 30%⁸, a figure that covers only big manufacturers.

The OECD warns that cross-country comparisons “should be made with great caution” because surveys define and measure AI use differently⁹. So read the table below as a range, not a league table.

Why do some surveys say most businesses already use AI?

Because they ask different questions, sometimes on behalf of companies that sell AI. Europe and the UK show it clearly. Eurostat, the EU’s official statistics office, found 20.0% of EU enterprises used AI in 2025³, while a study commissioned by Amazon Web Services (AWS) reported that 54% of European businesses use AI¹¹. In the UK, the government’s figure is 27% of businesses that handle digital data⁴, while AWS’s latest UK study reports 64% of organisations¹².

Neither is necessarily wrong. We don’t know every detail of how each survey was run, but the gap shows how much the question matters: counting any use of an AI tool gives a far higher figure than asking whether AI is part of how the business works. Before trusting a headline figure, ask who was asked, what counted as “using AI” and who paid for the survey.

How fast is AI adoption growing?

Fast. In Canada, the share of businesses using AI went from 6.1% in 2024 to 12.2% in 2025 and 19.2% in 2026². In the EU, it rose by 6.47 percentage points in a single year, to 19.95% in 2025¹³. Commissioned studies point the same way: AWS’s European figure went from 33% to 54% in two years¹¹.

To make that concrete, we extended each official trend in a straight line. If Canada and the EU simply kept their recent pace, around half of EU businesses, and well over a third of Canadian ones, would be using AI by 2030. That’s an illustration, not a forecast: real adoption usually speeds up and then levels off. But it shows how much will change within five years, and how quickly “most businesses haven’t started” stops being true.

How many small businesses use AI, and are they falling behind?

Fewer than large ones, and often far fewer, in every market we found that reports it by size. In the EU, 17% of small enterprises used AI in 2025, against 55% of large ones¹³. In the US, 37% of firms with at least 250 employees use AI, while less than 20% of firms with four or fewer do¹. Canada’s gap is narrower: 19.9% of businesses with one to four employees against 27.8% of those with 100 or more². In South Korea it’s 27.4% of firms with 50 to 249 employees against 63.3% of those with more than 250¹⁴; in Singapore, 14.5% of smaller firms against 62.5% of larger ones¹⁰; and in Hong Kong, 17% against 52%⁶. Across the OECD as a whole, large firms are more than three times as likely to use AI as small ones⁹. The UK and China figures we found don’t split out small firms, so they’re not in the chart.

And the gap isn’t standing still. In the US, AI use grew among firms with at least 20 employees but didn’t change significantly among smaller ones¹. In Singapore, smaller firms gained about ten percentage points in a year while larger ones gained nearly twenty, from 44% to 62.5%¹⁰.

Why does the gap keep widening?

Because the businesses using AI well keep compounding their advantage. In Britain, the statistics office found that businesses using advanced digital technologies such as AI and cloud computing achieved around 19% higher turnover per worker, even after accounting for management practices and other differences⁴. In the US, Census researchers found a robust positive correlation between firms’ commercial performance and how broadly they use AI¹⁵.

The leaders are also speeding up. McKinsey’s AI “high performers” (the roughly 6% of companies that attribute at least 5% of their operating profit to AI) are changing how they work faster than everyone else: nearly three-quarters now fundamentally redesign workflows around AI, up from 55% a year earlier, while just one-quarter of other companies do¹⁶. Size compounds too. In the US, firms with 1,000 or more employees grew their share of employment from 37.6% in 2000 to 42.3% in 2025, and economists at the Peterson Institute think “further concentration seems likely from future AI adoption”¹⁷. The OECD warns that divides between leading and other firms “may widen”⁹.

So can small businesses afford to wait?

No. You’re not late yet, but every year you wait makes catching up harder. The Peterson Institute’s economists put it plainly: “faster adopters are likely to gain competitive advantages over their slower rivals”¹⁷.

Our view: each year a competitor runs on AI, it gets a little more efficient, wins a few more customers and learns more about what works. Those gains compound. Waiting doesn’t keep you level; it lets the gap grow, and catching up later means chasing a moving target with thinner margins and fewer customers to fund the effort. The good news is that in most markets the majority hasn’t moved yet. That’s your last window to move early. Start small, measure it, and start now.

Are people ahead of businesses?

Yes, and that raises the stakes. Stanford’s AI Index puts generative AI use at 61% of the population in Singapore and 28.3% in the US¹⁸. Japan’s government white paper found 58.8% of people had used generative AI in fiscal 2025¹⁹, and in China the adoption rate among people reached 42.8% by December 2025²⁰. These figures count people, not companies, and come from different surveys, so they can’t be compared exactly with the business figures. But the pattern is hard to miss: your staff and your customers are getting used to AI, whether or not your business has a plan for it.

Why do some businesses hold back?

In Canada and Australia, mostly relevance and trust. In Canada, 41.4% of businesses with one to four employees say AI isn’t relevant to them, against 21.3% of businesses with 100 or more employees². In Australia, 54% of small and medium businesses that don’t use AI say it isn’t relevant, and around 65% cite distrust of AI decision-making or a preference to keep human control⁵.

Both are answerable. On trust: keep a person in charge of anything that reaches a customer. In the US, 66% of businesses using AI use it only to help people with tasks¹⁵, and only 2% reported AI-related job cuts¹⁵. On relevance: start with one task where AI would move a number you already track. That’s the subject of our next issue: how to put a number on an AI project.