The Federal Reserve Is Now Tracking Which Firms Use AI — And the Gap Is Getting Harder to Close

This isn't a vendor survey. It isn't industry association polling. It is the Federal Reserve System formally tracking AI adoption as a macroeconomic variable, using Census Bureau Business Trends and Outlook Survey data. The kind of data that gets cited in congressional testimony and Federal Reserve Board meetings.

For a professional services firm owner who has been watching the AI conversation from the sidelines — evaluating tools, reading the news, hearing about it at conferences — this data marks a turning point. The gap between firms that have adopted AI and firms that haven't is now large enough to be tracked by the federal government. That's not a trend coming in 18 months. That's the current state of your competitive landscape.

---

The Numbers the Federal Reserve Just Published

The Federal Reserve's FEDS Note draws on the Census Bureau's Business Trends and Outlook Survey, which was expanded in November 2025 to include formal AI adoption tracking. The Fed is now monitoring AI the same way it monitors employment, capital investment, and output growth.

The core findings:

• 18% of US firms overall have adopted AI in some form
• ~75% of large firms (250+ employees) are using generative AI
• Small firms (under 50 employees) are in the high single digits — roughly 5–10%, with the task title figure of approximately 8% reflecting midpoint estimates
• The gap between large and small firm adoption is not closing — it is widening

A companion FEDS Note published the week before — "AI Adoption and Firms' Job-Posting Behavior" (March 27, 2026) — adds another layer: firms that have adopted AI are already changing the mix of roles they hire for. The labor market signals of AI adoption are visible. The firms at 75% adoption are not just working differently. They are restructuring what kind of human work they need.

What makes these numbers different from every AI vendor survey you've read:

• The source is neutral. The Federal Reserve has no interest in selling AI tools.
• The methodology is rigorous. Census Bureau survey data is the same infrastructure used to measure GDP, employment, and inflation.
• It's a baseline. This is the starting point for ongoing tracking. Future quarters will show the trajectory.

This is the data that validates what many professional services firm owners have sensed but not been able to quantify: the firms ahead of them aren't just experimenting with AI. They're operating at scale, and the gap is already large enough for the government to measure.

For more context on what the AI adoption gap means for professional services firms specifically, the hub page covers the full landscape across firm types.

---

Why the Gap Compounds

Here is the dynamic that makes the 75%-vs.-8% gap more serious than it appears at first glance: adoption compounds.

A firm that started using AI in January 2025 spent Q1 2025 getting one tool working in one workflow. By Q3 2025, they had refined that workflow, identified a second application, and started building institutional knowledge — internal documentation, trained staff, refined prompts, workflow integrations. By Q1 2026, they are operating with AI embedded in multiple workflows and the organizational knowledge to evaluate new tools quickly.

A firm that starts in April 2026 starts from zero. Not just technologically — organizationally. They have no internal knowledge base. No prompts. No trained staff. No embedded workflows. They're starting a learning curve that the early adopters completed 18 months ago.

This is why the Fed data matters beyond the headline number. It's not that 75% of large firms use AI and 8% of small firms do. It's that the 75% have had 18 months to refine their processes while the 8% are still evaluating.

The job-posting companion paper adds a concrete labor market angle: AI-adopting firms are changing what they hire for. They need fewer people doing routine tasks and more people doing judgment-intensive work — evaluation, strategy, client relationships, complex problem-solving. This matters to small professional services firms competing for the same talent pool. The firms at 75% adoption are training their hires differently and selecting for different skills. The gap is not just in workflows. It's in what the workforce looks like.

Every quarter of inaction does not just delay adoption. It increases the cost of the transition — more habits to change, more ground to cover, more client expectations already recalibrated by firms that moved faster.

---

What "Closing the Gap" Actually Looks Like at a 10-Person Firm

The framing of a 75%-vs.-8% gap can feel paralyzing. It's not. "Closing the gap" does not mean getting to 75% adoption. It means being measurably better than you were 90 days ago.

At a 10-person professional services firm, closing the gap looks like this:

Step 1: Pick one tool for one workflow and use it daily for 30 days.

The criteria: it should address something repetitive that happens at least 3x per week. Strong candidates:

• AI meeting notes (Otter.ai, Fireflies.ai, or your video platform's built-in tool) — meeting summaries written before the call ends
• AI-drafted client email responses — you edit, it drafts
• First-pass research summaries — context on a client's situation, an industry, or a regulatory issue, synthesized in 5 minutes instead of an hour

This is not a transformation. It's a 30-day test with one tool. The cost is typically under $50/month. The time investment is a few hours in week one to set it up, then daily use.

Step 2: Document what changed.

At the end of 30 days, answer three questions:

• How much time did this save per week?
• What errors did it catch, or what quality improvements did you notice?
• Would you stop using it if you had to?

This is your internal ROI signal. It tells you whether the tool is worth expanding. It also gives you something concrete to say when staff ask why you're adopting AI, or when clients ask how you use it.

Step 3: Train one other person.

This is the step most firms skip, and it's the most important one. An AI workflow used by one person is a personal productivity tool. An AI workflow used by two people is institutional knowledge. Three people and it's a process. What separates a pilot from adoption is whether the knowledge transfers.

Train one team member before you evaluate whether to expand. Have them use the same tool in the same workflow for two weeks. Then compare notes. You'll surface edge cases, refine the workflow, and build the organizational muscle you need for step four, five, and six.

The goal is not to get from 8% to 75% in one quarter. The goal is to be one step ahead of where you were — and then to do that again.

For a practical look at which AI tools make sense for firms of your size and type, the tools guide covers the specific options with implementation notes.

---

Why Small Firms Have an Advantage the Data Doesn't Show

The 75% adoption rate at large firms is real. But there's a structural reality that number obscures.

Large firms implementing AI at scale face:

• Integration complexity — legacy systems, custom software, and enterprise software contracts that weren't built for AI
• Security review and procurement processes — a new tool at a 500-person firm may require 6 months of legal review, IT approval, and vendor vetting
• Change management at scale — getting 300 people to change a workflow is a multi-quarter initiative with training, internal communication, and enforcement mechanisms
• Institutional inertia — the same size that creates resources also creates bureaucracy

A 10-person professional services firm can make a decision in a week. Implement it in a day. Adjust it immediately when it doesn't work. If the tool isn't delivering value after 30 days, you cancel the subscription and try a different one. That decision cycle would take 6 months at a large firm.

The structural agility advantage is real — but only if you use it.

The firms that will look back on 2026 as the year they made the crossing are not the ones that had the biggest AI budgets. They're the ones that made a decision, started, and learned from doing rather than evaluating.

---

What to Do This Week

The Federal Reserve's data is not a call to panic. It's a call to start. Here are three specific actions you can take this week:

1. Read the FEDS Note itself. The Federal Reserve FEDS Note on AI adoption (April 3, 2026) is 8 pages. It takes 20 minutes to read. If you're going to make decisions about AI adoption for your firm, the baseline data should come from the primary source — not a summary of a summary. The companion note on AI adoption and job-posting behavior (March 27, 2026) is worth reading immediately after if you're thinking about hiring.

2. Identify one repetitive task your team does in client delivery that takes 30+ minutes per week. Be specific. Not "research" — "summarizing the client's prior year tax situation before every planning call." Not "communication" — "drafting the weekly status update email to our three largest clients." Write it down. This is the workflow you're going to test.

3. Find and trial one AI tool for that task this week. Most professional AI tools have a free trial. Start the trial, apply it to the specific workflow you identified, and use it at least three times before Friday. You're not committing to anything. You're collecting data on whether it works.

---

That's the minimum viable response to a government report confirming that 75% of large firms are operating with capabilities you don't currently have. Not a strategy retreat. Not a technology committee. One tool. One workflow. Three uses by Friday.

The gap is real. The compounding is real. The advantage of acting now — while the early-mover position in your specific market is still available — is also real.

For a complete framework on building an AI policy that covers client data, vendor vetting, and staff governance alongside your adoption work, see the professional services AI policy template.

---

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, tool comparisons, and implementation guides — including this week's complete AI tool evaluation checklist by firm type.

Subscribe to The Crossing Report →