What does the Census BTOS survey measure about AI?

The BTOS measures whether businesses are using AI in business functions — operations, production, customer service, and similar workflows — not whether they have subscribed to an AI tool or experimented informally. It is a narrower, harder definition of AI use than surveys that ask 'have you tried AI,' which is why the professional services 62% figure carries more weight than softer adoption surveys.

Are small professional services firms behind on AI adoption?

Not necessarily. The BTOS data shows a U-shape adoption distribution: the smallest firms (1–4 employees) have the second-highest adoption rate among all firm size categories. However, 5–50 person firms — the typical professional services practice — show slower adoption than both the smallest and largest firms. The gap within professional services is between firms that have adopted AI in production workflows and those still experimenting informally.

What is the performance gap between AI-adopting and non-adopting firms?

Among AI-adopting firms, 23.1% of employment is at companies rating their performance as 'excellent,' versus 13.8% at non-adopting firms — a 9.3-percentage-point gap. This correlation suggests firms using AI are achieving higher performance ratings on average, though the Census data does not establish causation.

How does Census BTOS data compare to Federal Reserve small business AI data?

They measure different things. The SF Fed 'Early Findings' report (March 2026) found 46% of small firms 'use AI' — a broader definition including any AI tool use. Census BTOS uses a narrower definition (AI in production and business functions) and finds 17.3% nationally. Within professional services, the 62% BTOS figure likely reflects higher formal adoption in client-facing workflows — document drafting, research, analysis — rather than casual use.

62% of Professional Services Use AI (2026)

Q: What percentage of professional services firms use generative AI?

62% as of the US Census Bureau's Business Trends and Outlook Survey (BTOS), released April 23, 2026. Professional services is the second-highest adopting sector, just below financial services (63%), and well above the national business average of 17.3%.

Published: May 2026 | By: The Crossing Report

Most industry AI surveys come from vendors. The one released April 23, 2026 came from the US Census Bureau — and the number it found for your sector is worth stopping on.

62% of Professional, Scientific, and Technical Services firms use generative AI. That's the US Census Bureau's Business Trends and Outlook Survey (BTOS), April 2026 release. Professional services is the second-highest adopting sector in the US economy. Only financial services is higher, at 63%.

The national average? 17.3%.

If you assumed your peers were still mostly watching from the sidelines, this data says otherwise. The professional services AI adoption rate in 2026 is not a laggard number — it's near the top of the economy. The question this piece answers: what does that actually mean for a 15-person accounting firm, law practice, or consulting shop?

What the Census Actually Measured (and Why It Matters)

The Census Bureau's BTOS is not a vendor survey. It is not a self-selected online poll of tech-forward readers. It is a federal business survey — the same infrastructure used to measure GDP components, manufacturing output, and employment trends. When the Census Bureau says 62% of professional services firms use generative AI in business functions, that is a harder measure than almost anything else you will find.

The definition matters here. The BTOS measures AI use in business functions — operations, production, client service, and similar active workflows. It does not count someone who tried ChatGPT once and abandoned it. It is not counting employees who use a consumer AI app at home. The threshold is: is your firm actively using generative AI in how it operates?

By that standard — a real, functional definition — 62% of professional services firms say yes.

That is the number your competitors are answering when they say yes to a client asking "does your firm use AI?" It is the number behind the firms that are turning around deliverables faster, taking on more work without additional staff, and showing up to business development conversations with AI as a differentiator rather than a liability.

The BTOS release date is April 23, 2026. Source: US Census Bureau Business Trends and Outlook Survey.

The Professional Services Number That Should Get Your Attention

Here are the sector comparisons from the April 2026 BTOS:

Financial services: 63% generative AI adoption
Professional, Scientific, and Technical Services: 62%
National average (all businesses): 17.3%
Projected national average by mid-2026: 22%

Professional services is not an outlier laggard. It is running at nearly 3.5x the national average. The image many firm owners carry — of a sector that moves slowly on technology, prefers relationships over tools — does not match what the Census data is measuring.

The practical implication: when you are competing for a client against another firm, the probability that the other firm uses AI in how it delivers its work is now more than 60%. That is the baseline you are competing against. Not the baseline from three years ago, when AI tools were curiosities. The baseline today is a majority of your direct competitors using these tools in production.

For accounting firms specifically, that is not a future threat. It is the current condition.

The Firm Size Finding That Surprises Most People

The BTOS found a U-shaped distribution in AI adoption across firm sizes. The largest firms (500+ employees) have the highest adoption, as expected. But the second-highest? Firms with 1–4 employees.

The mid-size firms — 5–99 employees, which is where most professional services practices sit — show slower adoption than both the smallest and largest categories.

This matters for two reasons.

First, the solo practitioner and micro-firm market is not waiting for enterprise-grade tools before moving. They are moving now, with the same tools you have access to. A one-person CPA shop can adopt AI faster than a 25-person firm because there is no change management problem — it is just the owner deciding to use a new workflow.

Second, the 5–50 person firm range — the core of The Crossing Report's audience — is in a squeeze. You are being pressured from below by micro-firms moving faster than expected, and from above by large firms with dedicated AI implementation budgets. The firms that will differentiate themselves in this size range are the ones that treat AI adoption as a competitive positioning decision, not a technology experiment.

The U-shape is not a curiosity. It is a competitive map. If you are in the 5–50 person range, you are in the slowest-moving tier, which means the window for differentiation by moving faster than your immediate peer group is still open — but it is narrowing.

The Performance Gap You're Paying For If You Haven't Started

The BTOS found a clear performance correlation between AI-adopting and non-adopting firms:

AI-adopting firms: 23.1% of employment is at firms rating their business performance as "excellent"
Non-adopting firms: 13.8%
The gap: 9.3 percentage points

The Census data is correlational, not causal — firms that are generally better-managed may both adopt AI and perform better for unrelated reasons. But the direction and size of the gap are consistent with what firm owners who have integrated AI report: better client delivery speed, higher capacity without headcount growth, and fewer bottlenecks in the workflows that are most client-visible.

For a professional services firm where client retention depends on perceived quality and responsiveness, a 9.3-point gap in "excellent" performance ratings is not small. It maps directly to the difference between clients who renew without asking for a discount and clients who shop around.

The performance gap also compounds. Firms that started integrating AI in 2024 have two years of workflow refinement behind them. Their review cycles are faster, their templates are more refined, their staff are more practiced. The gap between those firms and firms starting now is not just current capability — it is organizational learning that is hard to replicate quickly.

For more on how to measure whether AI is actually improving your firm's financial performance, the AI ROI measurement framework for professional services firms covers the specific metrics worth tracking.

Where Adoption Meets Integration — The Gap That Matters

The 62% professional services figure is adoption, not integration. Those two things are not the same.

The Federal Reserve Bank of San Francisco's March 2026 data found that while 46% of small firms use AI broadly, only 7% have fully integrated it into their business operations. The SF Fed and Census Bureau are measuring different things — the SF Fed uses a broader definition of "uses AI," while Census BTOS uses a narrower production-use threshold. But the underlying story is the same: most firms using AI are not getting systemic benefit from it.

Adoption means someone at your firm uses an AI tool regularly. Integration means AI is embedded in the workflows that determine how much work you can take on, how fast you deliver it, and how consistently you deliver at quality.

The difference in outcome between adoption and integration:

Adoption captures: time savings on individual tasks (drafting, summarizing, researching), productivity improvement for the people using the tools, reduced grind on lower-value work.

Integration captures: the ability to take on more clients without adding headcount, faster client delivery that compounds into higher satisfaction and retention, and AI-enabled services that were not possible without it.

The 62% professional services adoption number is real. But the number that actually determines competitive position is the smaller subset of those firms that have moved from "our staff uses AI tools" to "our client workflows run on AI."

If you want a breakdown of how accounting firms specifically are bridging that gap, the four-strategy framework for closing the accounting firm AI adoption gap covers the specific transition each stage requires.

Three Questions to Answer This Week

The Census data benchmarks where your sector is. These three questions benchmark where your firm is:

1. Which functions are you using AI in right now? Be specific. Not "we use ChatGPT sometimes." Name the workflows: client intake drafts, engagement letters, meeting summaries, research memos, deliverable templates. If you cannot name at least two workflows where AI is used consistently by more than one person, your adoption is at the experimenter stage — not production use.

2. How many client-facing workflows use AI? This is the integration question. Internal productivity gains are real but secondary. The leverage is in client-facing work — the deliverables, communications, and services clients see and evaluate. If AI is only touching internal work, you are leaving the competitive advantage on the table.

3. Are you measuring any outcome? Hours saved, turnaround time, client satisfaction, capacity per employee. Pick one metric. Measure it before and after. If you cannot show a difference, you either have not integrated deeply enough or you are measuring the wrong thing.

Firms that have answered all three questions — and can defend the answers with specifics — are in the 62%. Firms that are integrating systematically and measuring outcomes are in the subset pulling ahead of it.

The Crossing Report Verdict

The Census BTOS data removes any remaining ambiguity about whether professional services has meaningfully adopted AI. At 62%, the sector is near the top of the US economy — not waiting, not watching, not deliberating.

The 17.3% national average is the context that makes that number land. Your sector is not average. Your peers are not average. The standard your clients will use to evaluate you is not the national average — it is the firm in your practice area that started integrating AI eighteen months ago.

The adoption window is still open. The integration window is where the advantage will be decided.

What to do this week:

Pick the single workflow at your firm that touches every client — intake, deliverable drafting, or client communication. Write down every step from start to finish. Mark the steps where someone starts from a blank page or transfers data manually. Those three or four steps are your AI integration points.

That is not a technology project. It is a decision about which workflow gets rebuilt first. Make it this week.

The Crossing Report is a weekly intelligence newsletter for professional services firm owners navigating the AI transition. Free subscribers get the top three insights each week. Premium subscribers get the full analysis, implementation guides, and tool comparisons.

Subscribe to The Crossing Report →

62% of Professional Services Firms Now Use Generative AI. What the Census Data Means for Yours.