Thomson Reuters 2026 AI in Professional Services Report: What the Data Means for Small Firms

Published: June 2026 | By: The Crossing Report

---

The TR 2026 AI in Professional Services Report at a Glance

In early 2026, Thomson Reuters surveyed 1,500+ professionals across legal, tax, accounting, audit, risk, and government services. The result is the most comprehensive benchmark available for AI adoption across the professional services industry — and the numbers are different from what most trade press headlines suggest.

The key findings:

• 40% of professional services firms now use GenAI — up from 22% last year. That's an 82% year-over-year increase. Not a slow curve.
• 80%+ of current GenAI users engage with it weekly. This isn't casual experimentation — it's becoming infrastructure.
• 66% of professionals feel optimistic about AI's future role in their work.
• Only 18% of organizations track AI ROI. An additional 40% don't know if their organization measures ROI at all.
• 15% of firms currently use agentic AI. Another 53% are planning or considering it.
• 77% expect agentic AI to be central to their workflow by 2030.
• GenAI could create $32 billion in savings and 240 hours per professional annually in the legal sector alone, based on current adoption trajectory.
• 26% of advisory and consulting work is now being delivered with AI assistance.
• 1 in 3 firms is seeing pricing pressure on hourly-rate models.
• Firms with strategic AI adoption achieve 3x ROI compared to firms that adopt tools without a strategy.

The gap in all of this coverage: Thomson Reuters' own write-up targets enterprise legal departments and Big Four accounting firms. The data actually tells a more urgent story for small firms.

---

What "40% Adoption" Means for Your Firm

The 22%→40% jump happened in 12 months. In 2025, roughly 1 in 5 professional services organizations had formalized AI use at the organization level. Today it's 2 in 5.

That shift matters to a small firm owner in a specific way: you are no longer in the early majority cohort if you haven't started yet. You're in the late majority.

Here's how the TR 2026 data maps to real firm positions:

If you're in the 40% using GenAI: You're mainstream. The urgency question is no longer whether to adopt — it's whether you're measuring anything. If you're using AI but not tracking whether it's actually saving you time or improving client output, you're in the same position as the 82% who fly blind. You're just flying blind with a tool.

If you're in the 60% not yet using GenAI: The 19% "no plans" figure is shrinking fast. The professional services market is not bifurcating — it's converging on AI adoption. Firms that hold out are not maintaining competitive parity. They're falling behind a median that moved in one year.

The St. Louis Federal Reserve's parallel research suggests these adoption numbers may undercount real deployment by as much as 18% — because many professionals use consumer AI tools (ChatGPT, Claude.ai) that never show up in enterprise surveys. Actual peer adoption is likely higher than the TR headline.

---

The ROI Gap Is the Real Strategic Problem

The headline finding in the TR 2026 report isn't the adoption rate. It's this: only 18% of professional services organizations track whether AI is generating value.

That means 82% of firms — many of them actively using AI right now — have no consistent way to know if their investment is working.

The breakdown is striking:

• 18% track ROI — mostly through internal metrics only
• 40% don't know if their organization measures ROI at all
• The remaining firms have a process that's either inconsistent or unmeasured

This isn't a small-firm problem. Large firms fail at this too. The TR data spans all firm sizes.

The measurement gap creates three specific risks for a small firm:

1. You can't justify continued AI investment. Partners and principals ask: "Is this actually worth what we're spending?" Without measurement, you have a story. You don't have an answer.

2. You can't compare your performance to competitors who are measuring. The firms in the 18% are making better second and third AI decisions because they know what their first decisions produced. You're iterating blind.

3. You can't identify which tools are working and which are sunk cost. Most firms that have been using AI for 18+ months are running 3–5 tools simultaneously. Without measurement, they don't know which one is doing the work.

For legal firms specifically, the Harvey LAB legal agent benchmark provides the first domain-specific measurement framework — a way to evaluate AI performance on real legal tasks, by practice area, with attorney-written criteria. See the Harvey LAB analysis here.

The ROI gap is not technically complex to close. It requires picking one workflow, establishing a baseline (time spent, error rate, or client throughput), and measuring for 30 days. The firms doing this are not more sophisticated than you. They just started earlier.

---

Agentic AI: The 2026–2030 Planning Window

The most important forward-looking finding in the TR 2026 report isn't about where firms are today. It's the gap between 15% current agentic AI use and 77% expected centrality by 2030.

That's a 62-percentage-point shift expected in four years. The professional services industry is telling itself — through a survey of 1,500+ practitioners — that within four years, AI that can execute multi-step tasks autonomously will be how the majority of firms operate.

What "agentic AI" means in practice:

Regular GenAI: you ask a question, you get an answer, you do something with it. Every step requires your input.

Agentic AI: you define a task with parameters, the AI executes across multiple steps — research, synthesis, drafting, cross-referencing, flagging — and presents you a completed output to review. One instruction replaces what used to be 8–12 sequential decisions on your part.

Today, 15% of professional services firms operate this way for at least one workflow. By 2030, 77% expect to.

Thomson Reuters' own leadership has made the agentic bet explicit: the prediction of one AI agent per professional — from Elizabeth Beastrom, president of tax and accounting professionals at Thomson Reuters — frames the 2030 window as infrastructure, not experimentation.

The planning implication for small firms:

Firms that build agentic AI foundations in 2026–2027 will have measurable operational advantages by 2028–2030. The firms waiting for the technology to mature before investing in workflow redesign will arrive at an agentic AI market where competitors have two to three years of institutional knowledge they don't.

The 2026–2027 window is not the time to deploy agentic AI everywhere. It's the time to identify your top three highest-volume, most-repetitive workflow categories and determine whether an agentic AI solution exists for each one. The deployment decision can come later. The identification work has to happen now.

---

What the Business Model Numbers Mean

Two findings in the TR 2026 report don't get enough coverage:

26% of advisory and consulting work is now AI-assisted. One in four professional services engagements involves an AI tool doing part of the work. The client experience is changing whether or not the firm has decided to change it.

1 in 3 firms is experiencing pricing pressure on hourly rate models. This doesn't mean hourly billing is dead — it means clients are increasingly aware that AI is compressing the time required for work they used to pay for by the hour. The firms that haven't addressed this in their client conversations yet will face it reactively, in a negotiation they didn't initiate.

The Wolters Kluwer 2026 data corroborates the TR finding: the percentage of professional services firms that see AI as a competitive differentiator jumped from 9% to 41% in 12 months. That's the industry waking up to the fact that AI isn't just an internal efficiency tool — it's becoming visible to clients as a differentiator between firms.

The strategy ROI gap is the clearest number in the report: firms that have a documented AI adoption strategy achieve 3x ROI compared to firms that have simply purchased tools and deployed them without a framework. The difference isn't access to better AI. It's having a plan that connects AI use to client outcomes and firm performance.

---

Three Actions for Small Firms Based on This Data

The TR 2026 report doesn't recommend these actions explicitly — they're the interpretation. But they're the most direct translation of the data into what a 10-person firm can actually do.

Action 1 (if you're in the 40% using GenAI):

Build a simple ROI tracking sheet. For your four highest-volume AI-assisted task types, measure time spent per task for four weeks — both before AI and after. This is the minimum to join the 18% who are measuring. It doesn't require a consultant or a software platform. A spreadsheet works. The data you generate in four weeks will tell you more about your AI investment than 18 months of intuition.

Action 2 (if you're in the 60% not yet using GenAI):

The window to adopt without urgency has closed. Pick one task type — client communication drafts, research summaries, or document review — and run a 30-day pilot with one tool. Not five tools. One. The goal is to complete the feedback loop: run the task, review the output, measure the time difference, decide. You cannot make a good AI adoption decision in the abstract. You need one concrete experience to reason from.

Action 3 (regardless of where you are):

Identify your highest-volume, most-repetitive workflow category — the work that, if it could happen faster and without a senior staff member doing every step, would change your firm's capacity meaningfully. Then search for whether an agentic AI solution exists for that specific workflow. Not AI broadly. That specific workflow. The 2026–2030 window is the strategic planning period. 2028 is not too late to deploy — but it is too late to start the research.

---

How to Read the TR 2026 Data for Your Firm Type

The TR report surveys professionals across legal, tax, accounting, consulting, staffing, and risk. Not every finding applies equally to every firm type.

Law firms (5–50 attorneys): The 240 hours per professional annually projection applies most directly here — legal work is heavily time-billed, research-intensive, and documentation-heavy. The agentic AI question is urgent for litigation support, research memos, and contract review workflows. The Harvey LAB benchmark is the right tool for evaluating AI quality by practice area before committing.

Accounting and CPA firms: The 22%→40% GenAI adoption jump is occurring fastest in tax and accounting, where high-volume, deadline-driven work (tax prep, audit support, advisory) is the clearest candidate for AI assistance. The 3x strategy ROI finding hits hardest here — accounting firms that adopted AI as a tool without redesigning advisory workflows are leaving the premium on the table.

Consulting firms: The 26% advisory AI-assistance figure is the most important number for consultants. If 1 in 4 client engagements at peer firms already includes AI-generated analysis or synthesis, the question of whether to use AI has already been answered for the market. The consultant's competitive question is now: which parts of your engagement process does AI enhance versus which parts require your judgment specifically?

Marketing agencies and staffing firms: The business model pressure finding — 1 in 3 firms seeing pricing shifts — is most visible in project-based professional services. Clients are increasingly aware that AI compresses deliverable timelines. Firms that have not proactively addressed AI's role in their pricing model will face that conversation reactively.

---

The Bottom Line

The TR 2026 report is not a prediction. It's a measurement of where professional services firms are right now, and where the industry has told us it's heading.

40% adoption. 18% measurement. 77% expecting agentic AI central to work by 2030.

If you're in the 40% without measurement, the gap you need to close is not adoption — it's accountability. If you're in the 60%, the question isn't whether to start — it's what to start with.

The firms in the 18% who are measuring aren't smarter or better resourced. They started the feedback loop earlier. The gap closes the day you create a spreadsheet and time one workflow.

---

Premium subscribers to The Crossing Report get the complete AI ROI measurement framework — the exact metrics the top 18% of professional services firms are using, and a 4-week tracking template you can implement before the next billing cycle.

Subscribe to The Crossing Report for weekly intelligence on AI and professional services firms.