The 91% AI Value Gap: Why Most Organizations Fall Short — and How You Don't (2026)

TL;DR. Thomson Reuters' 2026 Future of Professionals Report (1,816 professionals, 62 countries) finds AI adoption is no longer the problem — 74% of professionals use AI several times a week — but 91% say their organization falls short of what the technology could deliver. The gap has real costs: a quarter of frustrated professionals are considering leaving, and clients notice the shortfall. The opportunity hiding in that number: if nearly every organization is underdelivering, the individual professional who closes the gap personally becomes rare — and visibly valuable.

For the past three years, the AI conversation at work was about adoption: who's using it, who's allowed to, who's behind. That conversation is over. The new question is harder — and for individual professionals, more interesting.

The gap, in numbers

Thomson Reuters' fourth annual Future of Professionals Report, surveying 1,816 professionals across law, tax, accounting, audit, compliance, risk, and global trade in 62 countries (fielded March–April 2026), draws a picture that will feel familiar from the inside:

• 74% of professionals now use AI tools several times a week; 44% use them multiple times a day. Adoption is mainstream, not experimental.
• Yet 91% say their organization is falling short of what the technology could deliver — the number the report labels the "AI value gap."
• More than a third admit to unsanctioned AI use — tools their organization hasn't approved, or approved tools used in ways it can't see — mostly out of frustration with what's officially on offer.
• Clients feel it too: 78% of clients say AI-enabled quality improvements are essential, but only 6% say they consistently receive them.

Everyone is using AI. Almost no one thinks their organization is getting what it should out of it. Both things are true at once, and the second doesn't fix itself with more usage.

Why the gap exists

The report's diagnosis — echoed across its accompanying analyses — is that the bottleneck moved. It used to be access and willingness. Now it's strategy, training, and workflow design: organizations bought tools and granted licenses, but didn't decide which work should change, didn't train people past the login screen, and didn't redesign any process around what the tools can actually do.

That failure mode shows up at the individual level as a familiar day: you use AI constantly, in small ways, on top of a workflow designed in 2019. Drafts get faster; nothing structural changes; the promised transformation never quite arrives. Multiply that by an entire organization and you get 91%.

It also shows up as shadow usage. When a third of professionals route around sanctioned tools, that isn't rebellion — it's a signal that the official setup lags what people already know is possible. (It's also a real risk: unsanctioned tools mean nobody is checking where client data goes. If that's your current workaround, at minimum know what the tools do with your data.)

The costs are no longer hypothetical

Three findings turn the value gap from an abstraction into a line item:

• Talent walks. Among professionals reporting a value gap, 24% are considering leaving within two years — 13% within 12 months — at an estimated $232,000 replacement cost per professional. Skilled people increasingly treat a coherent AI environment as part of the job offer.
• Clients notice. The 78%-want-it / 6%-get-it spread is a quiet competitive verdict. The report estimates up to $143 billion in US client revenue at risk as clients drift toward firms that deliver AI-enabled quality and speed.
• The frustrated are the capable. The professionals most bothered by the gap are the ones who know what the tools can do — the same people organizations can least afford to lose.

If you recognize your own organization here, the natural reaction is frustration. The more useful reaction is arithmetic.

The individual opportunity inside a 91% statistic

Here's the reframe this report earns: if 91% of organizations are underdelivering on AI, then the professional who personally closes the gap is operating in the top decile of a skill almost every employer needs and almost none has institutionalized. You don't need your organization to transform to benefit. You need to be the exception inside it — or the obvious hire outside it.

What that looks like in practice:

1. Build a real personal workflow, not scattered usage. Frequency isn't proficiency — 74% of the surveyed professionals use AI weekly and still sit inside the gap. Pick the two or three recurring tasks that eat most of your week and build a deliberate, repeatable AI-assisted process for them, on tools your organization sanctions.
2. Get systematically good at review. The professionals who extract real value aren't the fastest prompters; they're the best editors of AI output. A structured evaluation checklist turns "I glance at it" into a repeatable quality step — and protects you from shipping the errors that make cautious colleagues swear off the tools.
3. Manage the supervision cost. Heavy AI use has a real mental price — the babysitting fatigue we covered in why AI makes you more productive and more tired. Batching review and setting quality gates is what makes an AI workflow sustainable rather than a sprint you quietly abandon.
4. Document your delta. Keep an unglamorous log: task, time before, time after, quality outcome. "AI saves me time" is an opinion; "this deliverable went from four hours to ninety minutes at the same quality bar" is evidence — for your review, your rate, or your next interview.
5. Become the local translator. In a gap organization, the person who can show a colleague a working pattern — not evangelize, just demonstrate — accumulates visibility fast. The report's talent findings cut both ways: organizations are losing AI-capable people and desperately trying to keep them.

One caution while you do this: being the exception doesn't mean going rogue. The unsanctioned-usage statistic is a warning label, not a playbook — build your edge on tools you're allowed to use, with client data handled the way your profession requires.

Where to start: find your own gap first

The 91% is an organizational number, but the same question applies one level down: where are you falling short of what the technology could deliver in your specific role? Most professionals can't answer precisely — which tasks to automate first, what to keep manual, what to learn next.

That's the question our free AI Readiness Audit is built for: about five minutes, and you get a profession-specific score plus a concrete picture of which parts of your workflow have the most value sitting on the table. It won't fix your organization's strategy — but it will tell you where your own gap is, which is the one you can actually close this quarter.

The honest caveats

The Future of Professionals Report surveys knowledge professionals in law, tax, accounting, compliance, and adjacent fields — the pattern likely rhymes elsewhere, but the exact numbers are that population's. "Falling short of potential" is also a perception measure: it captures what professionals believe their organization could achieve, not an audited benchmark. And the $143 billion revenue-at-risk figure is a modeled estimate, not an observed loss. None of that changes the core, well-supported finding: routine AI use is now nearly universal among professionals, and satisfaction with what organizations do with it is nearly nonexistent.

Gaps that wide don't persist forever. The professionals who close theirs early get to spend the next few years being the person everyone else is trying to catch.

Sources

• Thomson Reuters Institute: Future of Professionals Report 2026
• Thomson Reuters Institute: Future of Professionals 2026: As AI adoption grows, so do the challenges
• Thomson Reuters press release: AI is Ready but Firms are Not
• LawSites: Thomson Reuters 'Future of Professionals' Report Warns of Widening Gap Between AI Adoption and AI Value