What the 2026 Thomson Reuters AI in Professional Services Report Reveals for Tax and Audit Firms

The headline finding from Thomson Reuters' 2026 AI in Professional Services Report is hard to ignore: organization-wide AI adoption has nearly doubled to 40% in the past year, and 87% of professionals expect AI to be central to their workflow within five years. But the number that tells the more honest story is buried a few pages deeper. Only 18% of organizations are measuring whether their AI investments are actually working, and of those, 77% track only internal metrics like cost savings and employee usage. Almost none are measuring client satisfaction or revenue impact.

Having worked with tax and audit firms through several cycles of technology adoption, I've seen that gap before. Firms move quickly to adopt tools, but the metrics that would confirm whether those tools are changing business outcomes either don't exist yet or live in a spreadsheet nobody owns.

The fragmentation problem hiding in plain sight

The most common AI use cases in tax and audit — research, document summarization, drafting — are genuinely useful. They save time on repeatable tasks. But they operate in isolation from the systems that actually run the business: resource scheduling, project margins, WIP, revenue recognition.

This is what the Thomson Reuters report calls "fragmentation debt," and it explains a dynamic I see consistently in the field. The recent Impact of AI on Professional Services study from SPI Research corroborates it: data quality has ranked as the single greatest barrier to AI adoption for two consecutive years, rated a top concern. When underlying data is fragmented, AI has nowhere solid to stand, and firms end up stuck in lightweight, document-adjacent use cases rather than operational ones.

The result is a coordination burden that absorbs whatever efficiency gains the tools created. Thomson Reuters frames it plainly: "the operational impact of AI continues to be largely divorced from the business impact of AI."

What agentic AI actually looks like in a tax & audit context

The report draws a meaningful distinction between generative AI, which produces content, and agentic AI, which can reason through a problem and execute multi-step workflows independently. Only 15% of organizations currently use agentic AI tools, but 53% are already in the planning or consideration phase, and corporate tax departments are showing higher-than-average adoption rates. The next wave is closer than it looks.

The operational implications are concrete. A generative tool can summarize a regulatory update. An agentic system can scan a thousand documents for a specific compliance exposure, flag the relevant instances, route them to the right senior reviewer, and log the action, without anyone manually orchestrating that process. When pattern recognition and data validation happen systematically, the output becomes fully auditable in a way manual processes never quite achieve.

Getting there requires thinking less about deploying AI and more about redesigning how work moves through the organization. Human expertise concentrates where it creates the most value: complex risk judgment, client relationships, high-stakes advisory. The high-volume coordination runs on automated workflows.

The pressure isn't letting up

Corporate tax professionals are among the most enthusiastic AI adopters in the survey: only 3% said GenAI should not be applied to client work. The appetite is real. The infrastructure to act on it, in many firms, is still catching up.

Organizations with a formal AI strategy are more than three times more likely to realize positive ROI than those without one. Yet 82% of respondents say their organizations either aren't collecting AI ROI metrics or don't know whether they are. Most firms are running AI without a scorecard. And as SPI Research found, firms that embed AI deeply into their operations — across more than 80% of projects — achieve EBITDA of 23.8%, compared to just 10.2% among those where AI remains peripheral. The difference isn't which tools firms bought. It's how far into the business they let it operate.

The senior professionals who used to spend Monday mornings chasing status updates across disconnected systems should be spending that time on the work that actually justifies their billing rate. In the firms that have made that connection, the shift is already underway.

Most tax and audit firms have invested in AI. Fewer have invested in making it operational. That's the problem Certinia is built to solve.