Where Accounting Firms Are Actually Using GenAI

How Fast Has AI Spread Across Accounting Firms?

Two years is not a long time. In accounting, it's usually too short to change how a firm operates.

AI moved faster.

In 2024, only 8% of tax and accounting firms had built AI into their organization-wide workflows. A year later that number was 22%. In 2026 it hit 40%. That's not gradual adoption. It's a near-tripling in twenty-four months.

Thomson Reuters surveyed professionals across 27 countries for their 2026 AI in Professional Services Report. The picture wasn't a slow, cautious rollout. It was a profession that crossed a threshold and kept going.

What Tasks Are Accounting Firms Using GenAI For?

The 2025 Thomson Reuters survey of 1,700+ tax professionals showed where AI was landing inside the actual job. Six task categories. All above 50%.

Tax research led at 77%. Tax advisory — the part of the job that's supposed to require human judgment — came in at 62%. Document review was lowest at 53%. That spread was tight for a technology most firms hadn't deployed two years before.

The 2026 data doesn't re-run that task breakdown. What it shows is depth: 44% of firms that have adopted AI now use it daily. Another 29% use it weekly. The question shifted from "which tasks?" to "how embedded?"

If your firm still treats AI as an occasional research shortcut, you're behind where most of your peers have already moved.

What Is the Business Impact of Having an AI Strategy?

This is where the 2026 data gets specific in a way the earlier reports didn't.

Firms that have built AI into 80% or more of their projects average 23.8% EBITDA. Firms where AI is still peripheral average 10.2%. That's not a rounding difference. It's more than double the margin.

The 2025 Thomson Reuters data pointed to $52,000 in annual value per professional at firms with a real AI strategy. The 2026 data shows what that looks like at the firm level. The gap is measurable and it's growing.

Only 18% of firms currently track any AI ROI at all. Most of those track operational metrics — cost savings, hours reduced. Almost none measure client satisfaction or revenue impact. A firm that isn't measuring can't see the margin gap. A firm that can't see it won't feel pressure to close it.

Put it in dollar terms: at a 20-person CPA firm doing $4 million in annual revenue, the difference between the two EBITDA figures works out to roughly $546,000 in annual operating margin. That's a number firm partners can take to a budget conversation.

The firms that will feel this gap first are the ones competing for the same mid-market clients. When one firm can offer faster turnaround, cleaner work product, and lower friction — and tie that to a documented AI workflow — the comparison becomes visible to clients in a way it wasn't two years ago.

Is Agentic AI Already Coming to Accounting Firms?

Yes. Fifteen percent of accounting firms have already adopted agentic AI tools, and another 53% are actively planning for them.

Agentic AI refers to systems that don't just answer questions or draft text. They take multi-step actions on their own — running a workflow, flagging anomalies, updating a file, queuing a review for a partner. Fifteen percent of accounting firms have already adopted some form of it. Another 53% are actively planning or considering it.

That's 68% of the profession either in or moving toward agentic systems within the next year or two.

Ramp launched Stack in June 2026 as an early example of what purpose-built agentic AI looks like in accounting practice. It automates the monthly close, was tested on more than 200 accounting tasks by working accountants, and makes every decision reviewable. It's one data point in what will be a larger wave of tools built specifically for accounting workflows rather than adapted from general-purpose AI.

Firms building governance frameworks now — policies, review checkpoints, documented workflows — will have an easier transition. Firms that waited on basic AI policy will face the same gap again, one level up.

Why Do Most Accounting Firms Still Have No Formal AI Policy?

Because adoption moved faster than governance. Seventy-five percent of firms still have no written AI policy, even as most now use AI for client-facing work.

Wolters Kluwer's 2025 Future Ready Accountant Report found that 75% of accounting firms have no formal AI policy. That's close to the 70% Thomson Reuters found in early 2025. A year of rapid adoption didn't close the governance gap.

A staff member who pastes client tax data into ChatGPT counts as AI adoption. So does a firm with vetted tools, a documented review process, and a clear data handling standard. Both show up in the same adoption numbers.

Only one has a plan when something goes wrong.

Firms with written AI policies view the technology more positively — 84% compared to 44% at firms without one. That's not a coincidence. Policy turns scattered use into something you can build on and defend.

What Should Your Firm Check Before Next Tax Season?

If your firm uses AI in any capacity, three things are worth a look before the next busy season.

Write down your AI policy. One page is enough. It needs to say which tasks are AI-assisted, what review happens before client delivery, and how staff should handle client data in outside tools.

Find out what tools staff are actually using. Most firms learn their real AI stack through a short staff survey, not a memo. Ten minutes and a short form. What you find may change the conversation.

Check your E&O coverage. A growing number of malpractice carriers have updated their language around AI-assisted work. If you haven't asked your insurer about this in the past year, ask now.

Three-quarters of accounting firms still have no AI policy. A basic one-page document puts you ahead of most of your peers. The adoption gap is mostly closed. The governance gap is still wide open.