Field Notes
ReportNov 5, 2026 · 7 min read

Half of Professional Services Firms Your Size Use AI. Depth Is Where They Differ.

If you have been wondering whether your firm is behind on AI, the answer is that using it at all no longer tells anyone anything.

Greg JeffersonGreg JeffersonFounder, Everyday Creative Works

Exploring how AI is changing real work.

If you have been wondering whether your firm is behind on AI, the answer is that using it at all no longer tells anyone anything.

The Snapshot

The U.S. Census Bureau's Business Trends and Outlook Survey found that about half of professional services firms with 10 to 99 employees used AI in the past two weeks, in data collected October 5 to 18, 2026. [1][2][3] Among AI-using firms in the 2025-26 Census supplement, 57% used it in three or fewer business functions. [4] Adoption is now the baseline; depth is where firms differ. For a small professional services firm, that changes the question from whether you use AI to how far into the work it reaches.

Where your firm actually sits

The Census Bureau's Business Trends and Outlook Survey, or BTOS, put current AI use at 24.6% of U.S. businesses as of October 18, 2026, up from 17.3% on November 30, 2025. [5] By firm size on that date, BTOS recorded 26.0% for firms with 20 to 49 employees and 29.0% for firms with 50 to 99. [6] Every size class rose over that period, including the three bands under 20 employees. [6]

The national average is the wrong yardstick for a professional services firm, because BTOS put the whole sector at 43.8% on October 18, 2026. [7][2][8][3] Within that sector, use was 49.3% for firms with 10 to 19 employees, 47.8% for 20 to 49 and 51.6% for 50 to 99, in data collected October 5 to 18, 2026. [1][2][3] Those sector-by-size figures carry standard errors of 3.03 points for the 20 to 49 class and 4.97 points for the 50 to 99 class. [1][2][3]

Your peer group, not the national headline, is the number to measure yourself against, and in that group you are roughly at a coin flip.

A yes on a survey is a very low bar

The Census Bureau's core question asks only whether the business used AI in the past two weeks, and whether it expects to in the next six months. [9][10] Among adopting firms in the 2025-26 supplement, the most common functions were sales and marketing at 52%, strategy and business development at 45% and IT at 41%. [4] The same supplement found that 65% of firms limit worker use to three or fewer tasks. [4]

Theresa Dunne and Adam Scavette of the Federal Reserve Bank of Philadelphia reported that 57 percent of Third District AI adopters described their adoption as experimenting with the technology. [11] Just 3 percent of those adopters had fully integrated AI into their business operations. [11] Goldman Sachs says that in its survey of over 2,400 graduates of its 10,000 Small Businesses program, 86 percent reported using AI but only 15 percent said it was fully integrated into core operations. [12]

The evidence partly supports treating breadth of use, not the fact of use, as the thing worth tracking inside your firm.

Your staff may already be ahead of you

The Census working paper on AI diffusion found that 6.8% of firms reported worker-level AI use without any reported firm-level use. [4] Counting either kind of use puts aggregate AI incidence at nearly 24%, against a firm-level adoption rate of 18%. [4] The authors note that the firm and worker questions were likely answered by the same respondent, and that a respondent's lack of information on workers' practices can cause discrepancies. [4]

The same paper reports that in 23% of firms, workers use AI in work-related tasks, rising to 41% on an employment-weighted basis. [4] Most users, 66%, rely on AI solely to augment tasks, and AI-related employment decreases occur in only 2% of firms. [4]

If you have no AI policy, you may still have AI in the building, and the sensible first move is to find out what people already use.

Why the numbers you see elsewhere are so much higher

The authors of the working paper Firm Data on AI find that sample mix, weighting scheme and respondent role each move measured AI adoption by five to twelve percentage points. [13] They report that CEO and CFO respondents give adoption rates four to seven percentage points higher than other respondents. [13] Those factors compound, so a survey of small firms weighted by business counts and answered by non-executives reports substantially lower adoption. [13]

Jeffrey S. Allen of the Federal Reserve Board cautions that larger firms clear the bar for answering yes simply because of their scale and number of employees. [14] Allen adds that the relative impact on their business may not be as high. [14]

Do not let a headline adoption figure from another survey tell you that you are behind; it is often measuring a different thing entirely.

What to do with this

  • Benchmark against professional services firms of your own size rather than the all-business average.
  • Ask your team, individually, which tools they already use for work, because firm-level answers miss worker-level use.
  • Count how many business functions AI touches in your firm, since most adopters stop at three or fewer.
  • Plan for augmentation rather than headcount cuts, because that is what most AI-using firms report doing.
  • If you report progress upward or to clients, report depth of use, not a yes or no on adoption.

Limitations

The Census Bureau reports standard errors of 3.03 points for the 20 to 49 class and 4.97 points for the 50 to 99 class in the sector-by-size figures, so those estimates are not precise. [1][2][3] The survey's unit response rate was 12% nationally for the October 5 to 18, 2026 collection, and 14% for professional, scientific and technical services. [15][2][3] From October 19, 2026 the questions use the term Super Intelligence, and no data under that wording were available here, so it isn't yet clear whether answers shift. [5][16] Nothing in this research measures whether deeper use produces better results, so treat depth as a gap to close, not a proven payoff.

Sources

Work with ECW