What AI is actually doing inside small professional-services firms in 2026

A couple of years ago, the question in most small firms was "Should we even try this?" That's not the question anymore. Whether you run a law office, a CPA practice, an engineering shop or the land department at an oil and gas company, somebody on your team is already using AI. Often it's more than one person, and often nobody wrote down how.

I've spent 35+ years as a petroleum landman, and I've been a hands-on student of all things AI since 2023. What I see now is a shift from curiosity to habit. So let's look at what the research says firms are actually doing, and what it means for a firm your size.

How many professional-services firms are really using AI?

More than you'd think, and the numbers come from more than one direction. In the RSM Middle Market AI Survey 2026, 84% of the professional-services respondents said AI is fully or partially integrated into their operations. That's a small slice of the survey (50 firms), so treat it as a signal rather than a census.

A bigger study points the same way. SPI Research's 2026 study of 245 professional-services organizations found that 43.6% of projects now use AI, up from 23.1% in 2025.

Law is a good example. Clio's 2025 Legal Trends Report found that 79% of legal professionals use AI in their firms, and that 71% of solo firms report using it too. Small isn't the same as left out.

Not every profession is moving at the same speed, though. The AICPA's 2025 MAP Survey, reported in the Journal of Accountancy, found that only 13% of CPA firms said they'd already implemented AI and automation successfully, even though 65% had an open or proactive attitude. In architecture, the AIA's 2025 research found only 8% of firms had implemented AI solutions, with another 20% working on it (AIA, 2025). Interest is high everywhere. Follow-through varies.

What are firms actually using AI for?

Mostly the unglamorous stuff, and that's a good thing. RSM's survey found that the top goal for professional-services firms over the next year was employee productivity (56%). RSM principal Drew Faries put the job list plainly: AI should support "research, drafting, summarization, knowledge access and workflow acceleration."

In practice, that looks like:

• First drafts. Engagement letters, client emails, proposals and memo outlines that a person then fixes and finishes.
• Summaries. Long documents, call notes and email threads boiled down so someone can decide faster.
• Research starts. A quick map of a topic before the real research begins, with every source checked.
• Finding things. Asking questions of the firm's own procedures, templates and past work instead of digging through folders.

The tools are mostly general ones. Bluehost's 2026 study of 350 U.S. small-business owners found ChatGPT in 73% of small businesses, more than five times the share of any specialized AI tool (Bluehost, 2026). If you're wondering what to add after that, I wrote a separate piece on how to pick an AI tool for a small firm.

Where is AI still falling short?

Two places: the data and the money.

On data, RSM found that among professional-services firms whose pilots had only moderate or limited success, the most common causes were data quality (45%) and integration problems (41%). That was a small group of 29 firms, but it matches what most owners already know. If your files are scattered across five systems and three people's inboxes, AI won't magically sort that out.

On money, SPI calls it "the leak." Its headline finding is that 84.8% of firms say AI saves them time, but far fewer say it generates revenue. Saved hours that just disappear into the week don't show up on the bottom line.

Who is responsible when AI helps with client work?

You are. Every serious source says the same thing in different words. RSM says AI "should operate within a human-in-the-loop model," with people "standing behind the advice provided to clients." For lawyers, the American Bar Association's Formal Opinion 512 spells out duties of competence, confidentiality, client communication, supervision and reasonable fees when lawyers use generative AI. I go deeper on that in where AI helps a small law firm, and the CPA version is in should a CPA firm use AI on client work.

AI can draft it. A person still has to own it.

What should a small firm do with the time AI saves?

Decide on purpose. The AICPA MAP Survey asked CPA firms that question, and a third of them didn't have a plan yet. Among the answers, firms said they'd reduce hours for better work-life balance (45%), offer more advisory work (40%) or take on more clients without adding staff (39%). Any of those can be the right call. Drifting isn't.

A simple way to start: pick one recurring task, like monthly client update emails, and track how long it takes before and after. Then write down where that time goes. That one habit fixes most of SPI's "leak."

What does this mean for the records your firm already owns?

Here's the part that doesn't get talked about much. The same AI shift that's changing how firms work has also made company-owned work product more interesting. Years of proposals, procedures, support tickets and project files sit in everyday software at most established firms. Some of those records may qualify for licensing, and many won't. I lay out the difference in which business records may qualify for a license.

If you'd like to see what applies to your industry, start with the site built for it: BalanceSheetData for CPA firms, BlackGoldData for oil and gas companies and law firms, LicenseMyFiles for architecture, engineering and construction, and BidMyData for established businesses in general.

Where should you start this month?

1. Ask your team what they're already using AI for. You'll learn more in one honest conversation than in a vendor demo.
2. Write a one-page rule: what client information never goes into a public AI tool, and who reviews anything client-facing.
3. Pick one task to measure, and decide in advance what the saved time is for.
4. Choose tools from a ranked shortlist instead of chasing every new app.