Small business AI adoption: a compiled report

Small business AI adoption statistics show steady growth, but the level depends on who is asking and how. Government data from the Census Bureau puts business AI use near one in five. Business association surveys put small business AI use near two in three. Both can be accurate for what they measure.

This report puts the main sources side by side: the Census Bureau's Business Trends and Outlook Survey (BTOS), a Census working paper on early AI tracking, the Federal Reserve's April 2026 note comparing AI surveys, and the U.S. Chamber of Commerce's 2025 and 2026 small business technology reports. We explain why the numbers differ and what that means for a service business deciding where to start.

The Census Bureau's BTOS is a biweekly survey of U.S. businesses that has asked about AI use since fall 2023. A Census working paper by Bonney and colleagues, published in March 2024, reported that AI use rose from 3.7% to 5.4% between September 2023 and February 2024, and projected about 6.6% by early fall 2024. That early question asked about AI used to produce goods or services.

In November 2025 Census broadened the question to AI use in any business function. Under the new question, a Census story published in May 2026 reported that national AI use stayed between 17% and 20% from December 2025 to May 2026, reaching 19.8% as of May 3, 2026. Expected use over the next six months was 20% to 23%.

Yes. Larger firms use AI more, and the gap is growing. Census reported that AI use rose among firms with 20 or more employees between December 2025 and May 2026 but did not change significantly for firms with fewer than 20. Under 20% of the smallest firms, those with four or fewer employees, reported using AI.

The Federal Reserve's note found the same pattern in a second survey. In both the BTOS and the Survey of Business Uncertainty, the largest firms adopted AI the most. Under the revised Census question, the smallest firms looked similar to other firms under 20 employees and clearly below larger ones.

The Chamber's surveys report much higher use. Its 2025 small business technology report, produced with Teneo Research and published in August 2025, found 58% of small businesses self-identified as using generative AI, up from 40% in 2024 and 23% in 2023. State figures for using an AI platform ranged from 77% in Maine to 31% in West Virginia.

The 2026 edition reported 66% of small businesses use AI, up 8 points from 2025 and nearly three times the 2023 share. It also found 41% use six or more technology platforms and that 84% plan to increase technology use over the next two to three years. The Chamber's web summaries do not state sample size or fieldwork dates, which makes direct comparison harder.

Mostly because of who is sampled and how answers are weighted. Jeffrey Allen's FEDS Note for the Federal Reserve Board, published April 3, 2026, compared surveys from late 2025. The Census BTOS put firm AI use at 18% in December 2025. The Survey of Business Uncertainty put firm AI use at 78% when weighted by employment and 69.4% when each firm counted equally. A household survey, the Real-Time Population Survey, found about 41% of the labor force used generative AI for work.

Allen identifies sampling and weighting as the largest factor. The BTOS is weighted to match the population of firms, which is mostly very small businesses. Employment-weighted surveys count large firms more heavily, and large firms adopt AI at higher rates. Firms with 250 or more employees are 0.9% of firms but 56.2% of employment.

• Weighting: firm-weighted surveys (Census) read lower than employment-weighted ones (Survey of Business Uncertainty).
• Question wording: "AI to produce goods or services" reads lower than "AI in any business function" or "any AI-enabled tool".
• Who answers: about 10% to 11% of BTOS respondents said they did not know whether the business used AI.
• Pressure to report use: the Fed note raises the possibility that executives over-report AI use, and says more research is needed.
• Survey design: government surveys use large probability samples, while association surveys may draw from panels and publish less detail on method.

Growth is steady in the government data and faster in the business surveys. The Federal Reserve note found that, before Census changed its question in November 2025, BTOS adoption rose 68% over the prior year, an increase of 3.9 percentage points. About 9% of firms planned to use AI within six months in early 2025, and more than 20% expected to use it in the first half of 2026. Among workers, the Real-Time Population Survey found about 12% used generative AI at work daily.

The Chamber's surveys focus on effects inside the business. Its 2025 report found 82% of small businesses using AI had increased their workforce over the past year. In the 2026 report, 47% said AI is creating jobs at their business today, 6% said it is enabling headcount reductions, and 95% of AI-using small businesses said they are working to upskill employees, including 47% providing on-the-job training. These are self-reported and do not prove that AI caused the hiring, but they suggest most adopters see AI as a tool for their existing team rather than a replacement for it.

Use the Census figure as the baseline for how many firms actually use AI in their operations, and the Chamber figure as a signal of how many business leaders have tried AI tools in some form. If your team uses a chatbot to draft emails, the Chamber would likely count you as an AI user. The Census question asks whether the business uses AI in a business function, which more firms answer no to.

Either way, the trend is clear. Every source shows AI use rising, and both the Census and Fed data show larger firms moving faster. For a growing service company with 10 to 50 employees, that means competitors at the larger end of your market are more likely to be using AI in operations than the smallest shops are.

The adoption rate of other firms matters less than whether specific tasks in your business would run better with AI or automation. A property management company might start with after-hours maintenance requests, an insurance agency with quote follow-up and a physical therapy clinic with appointment reminders. Most of these are rule-based automation, with AI added only where it helps, such as drafting replies or summarizing calls.

Size also shapes the starting point. A five-person salon or law office usually gets the most from a few dependable automations, such as reminders, review requests and lead routing. A 40-person clinic or brokerage, closer to the firms Census shows adopting fastest, may also benefit from AI that drafts responses, summarizes calls or sorts incoming requests, with clear rules for staff on how to use it.

Monvell's Consulting work starts with a diagnostic of where time goes and which tasks repeat, then our Automation team builds only what has a clear purpose and a measurable result.

• List the repetitive tasks your team does each week and estimate the hours each takes.
• Separate tasks that follow clear rules (automation) from tasks that need judgment on unstructured text or speech (where AI can help).
• Start with one workflow where errors are low-risk and results are easy to measure.
• Write a short policy on which AI tools staff may use and what data they may not enter.
• Review results after 60 to 90 days before expanding.