A growing gap between artificial intelligence adoption and execution is putting client revenue and employee retention at risk for professional services firms, according to a new report from Thomson Reuters.
The company said its 2026 Future of Professionals report found that while AI tools are widely used across legal, tax, audit and risk professions, many organizations are failing to translate that usage into measurable business value, exposing them to financial and operational consequences.
The findings point to a disconnect that is beginning to affect both client relationships and workforce stability. Thomson Reuters estimates that up to US$143 billion in client revenue is at risk in the United States alone as firms struggle to meet expectations for AI-enabled services, while a significant share of professionals say they are considering leaving employers that fall short on AI delivery.
“We’re seeing a clear divide emerge,” said Steve Hasker, president and CEO of Thomson Reuters. “Firms that are operationalizing AI are pulling ahead. Those that aren’t are starting to take on real risk, across talent, clients, and financial performance. Closing that execution gap is now a business imperative for professional firms.”
The report is based on a global survey of more than 1,800 professionals conducted earlier this year across 62 countries, spanning private practice, corporate and government roles.
It found that AI adoption itself is not the primary issue. About 74 per cent of respondents said they use AI tools weekly, yet 91 per cent believe their organizations are not fully realizing the technology’s potential. This gap is contributing to unintended risks, including the rise of so-called “shadow AI” — the use of tools that have not been approved by employers.
Roughly one-third of lawyers, accountants and compliance professionals reported using unsanctioned AI tools, a figure that rises to 41 per cent among those who believe their organizations are moving too slowly on implementation. At the same time, respondents indicated high expectations for safeguards, with 96 per cent saying AI systems must protect confidential data, 94 per cent requiring verified, authoritative content and 90 per cent needing outputs they can explain and defend. However, 41 per cent said they lack access to tools that meet those standards.
The report also highlights a widening gap between strategy and execution inside organizations. While some firms have articulated AI ambitions, 35 per cent of respondents said those plans are not reflected in their day-to-day work, and nearly one in five said their organization still lacks a clear AI strategy.
That disconnect is increasingly influencing workforce decisions. One in four professionals — or 24 per cent — said they would consider leaving their employer within two years if they do not see the expected value from AI, with 13 per cent indicating they could leave within 12 months. Despite that, nearly half of senior leaders surveyed believe meaningful talent pressure remains at least three years away.
Access to AI tools is also emerging as a factor in recruitment. The report found that 62 per cent of respondents would consider the availability of professional-grade AI when evaluating a new role, and among those already using such tools, nearly one-third said they would decline a job offer that did not provide them.
Client expectations are shifting in parallel. According to the report, 78 per cent of corporate clients now consider AI-enabled quality improvements to be very important or essential, yet only six per cent believe most service providers are delivering on that expectation. As a result, 32 per cent of clients said they plan to reconsider provider relationships within the next 12 months, with some placing more than US$1 million in annual work under review.
“Not all AI is created equal. In professions where there is real liability, the standard has to be much higher,” said Hasker. “When outputs shape legal judgments, regulatory filings, or client advice, ‘almost right’ isn’t good enough. That’s why we build what we call Fiduciary-Grade AI, technology professionals can verify, trust, and ultimately stand behind.”
Thomson Reuters said the pressures outlined in the report are converging across risk management, talent retention and client demand, underscoring the need for firms to move beyond experimentation toward operational deployment of AI tools.
The company said the challenge is no longer technological readiness but execution, as organizations face increasing expectations for accountability in how AI is implemented and governed.
More from Retail Insider:
