Most owners want one answer from the AI statistics: are we behind? That depends on who did the counting, and on what they called “use”. Below are five figures, each with its source, its sample and its definition. Take them one at a time. Averaging them would blend questions that were never asked the same way.
17%: the size gap comes first
In 2025, 17% of small EU enterprises used AI, against 30.36% of medium ones and 55.03% of large ones. Eurostat counts firms with 10 to 49 staff as small, and “use” as at least one of eight technologies, from text mining and speech recognition to image generation and machine learning.
Singapore measures it its own way and finds the same shape. In the IMDA’s Singapore Digital Economy Report 2026, AI adoption among SMEs rose from 14.5% in 2024 to 23.4% in 2025, while non-SMEs reached 70.4%.
In the EU, a large company is more than three times as likely to use AI as a small one. You won’t see that gap in a list of software. You’ll see it in work a bigger competitor now does for less, like answering a quote request or writing a product page.
66%, or under one in five: it depends who asks
The US Chamber of Commerce’s 2026 report finds that 66% of US small businesses use AI. Teneo surveyed 3,732 firms with fewer than 250 employees between July 23 and August 6, 2026, and counted any AI platform. Generative AI chatbots lead, at 41%.
The US Census Bureau asks a different question in its Business Trends and Outlook Survey: is the business using AI “in any business function”? Between December 2025 and May 2026, between 17% and 20% of US employer businesses said yes. The share climbs with size, to 37% of firms with 250 employees or more.
Neither figure is wrong. One asks owners about any AI platform. The other is a national statistical survey whose wording changed in November 2025. A chatbot subscription can make a firm a user in one and not in the other. So don’t benchmark on “do we use AI”. Benchmark on which jobs run with it.
14%: plenty of trying, little running
In the Goldman Sachs 10,000 Small Businesses Voices survey, 76% of owners report using AI. Only 14% say it’s fully embedded in their core operations, and 73% say they’d benefit from more training and implementation resources. The sample is 1,256 alumni of the program, polled from January 27 to February 4, 2026: growth-minded owners, not a random cut of small business.
So even among owners who go looking for new tools, deep use is rare. The training figure points the same way: buying AI was the easy part, and setting it up is where they want help.
70.31%: what holds firms back
Eurostat also asked firms that had considered AI why they didn’t use it. In 2025, the most common reason was a lack of relevant expertise, at 70.31%. Next came unclear legal consequences, at 53.61%, and data protection concerns, at 52.72%.
All three are problems of knowledge and rules. An owner can close them without a large budget: one named person who learns the tools, a short written rule on what data goes into them, and a check of the legal duties where you sell.
43%: some owners already see it in revenue
In the QuickBooks 2026 AI Impact Report, 43% of US businesses say AI has increased their revenue, and 2% say it’s gone the other way. Intuit polled more than 34,000 owners in seven waves from July 2024 to January 2026, with University of Chicago economists. These are QuickBooks customers reporting on themselves, not an audit.
That’s a large minority seeing revenue, and a majority that doesn’t yet. If you’re in the second group, start by writing down what you’ll measure before the next subscription renews.
What to do with the five
- Compare jobs, not adoption rates. List the three tasks a larger competitor most likely runs with AI, and check yours.
- Run one job end to end. A chatbot on the side counts as use in a survey. It rarely moves your costs.
- Fix expertise first. Name an owner, train them, write the data rule.
- Write the measure before you spend. Hours saved, quotes sent or revenue per customer, recorded before and after.
Every figure here is dated, and most have a short shelf life. Eurostat’s 2025 figures were extracted in December 2025, and we’ll update this page when the 2026 edition appears. The Eurostat reasons were corrected on November 4, 2026: Eurostat’s article prints the 2024 values, and the 2025 values above come from its database. For the size gap in detail, read why Europe’s small firms lag on AI. Corrections to hello@bearingbridge.com.







