Organisations are struggling to translate AI usage into real value, finds study
Nearly one in four professionals who feel their employer is failing to deliver on AI are prepared to resign within two years — a retention threat that puts HR at the centre of what Thomson Reuters calls an AI "execution gap”..
The report concludes that the central risk to professional firms is no longer slow AI adoption but the failure to turn AI strategy into day-to-day practice.
It estimates that failing to execute on AI could put up to US$143 billion in U.S. client revenue at risk, along with skilled staff at an estimated US$232,000 per replacement.
Organisations are deploying AI into core operations faster than they can recruit and reskill the workers needed to run it, according to a previous report.
Talent risk tops the findings
Among professionals experiencing a gap between what AI can do and what their organisation delivers, 24% are weighing an exit within two years and 13% within 12 months, according to Thomson Reuters. More than 9 in 10 reported experiencing some degree of AI "value gap,” based on a survey of 1,816 professionals across law, tax, audit, accounting, compliance, risk and global trade in 62 countries, conducted in March and April 2026.
Usage is not the problem: 74% of professionals are already using AI tools every week, but organizations are struggling to translate that usage into real value.
The risk is concentrated among mid-career professionals, whom the report calls the most embedded AI users and the most operationally critical staff. Almost three in ten said they would change jobs within two years if AI fails to deliver.
The report also flags a pipeline cost: when senior staff leave, they take the mentorship junior colleagues depend on, and 71% said early-career roles need structured support to build the judgment AI risks displacing.
Shadow AI raises governance concerns
The report found that 34% of professionals use AI tools their organisation has not sanctioned and cannot monitor, rising to 41% among those who feel their employer is moving too slowly.
That use is not careless: 96% said their AI must safeguard confidential data, 94% require verified content and 90% need explainable outputs, yet 41% said they lack access to tools meeting those conditions.
Steve Hasker, president and CEO of Thomson Reuters, said the bar must be higher in regulated fields. "When outputs shape legal judgments, regulatory filings, or client advice, 'almost right' isn't good enough," he said.
Tool access shapes recruitment and retention
Access to professional-grade AI is now a hiring factor, Thomson Reuters said, with 62% reporting it would influence whether they accept a new role.
The pull is strongest among current users: nearly one in three would turn down a role that did not offer such tools, compared with 12% of those without access.
Clients are adding pressure. The report found 78% consider AI-enabled quality improvements very important or essential, but just 6% say most providers deliver, and 32% expect to reconsider relationships with lagging firms within 12 months.
Report frames the problem as change management
The core failure is organisational, not technical, the report said: 35% of those with a named AI strategy said it is invisible in their day-to-day work, and 17% said their organisation has no AI direction at all.
Asked why strategies stall, respondents cited tools not being in place (47%), inadequate training (43%), no clear priorities (32%) and no shared understanding (30%) — what Thomson Reuters called classic change-management problems.
Hasker said the human element remains decisive. "AI is a powerful force multiplier, but the judgment, relationships and accountability remain human, and that won't change," he said.
For those who are banking on AI training to equip their workers with the necessary skills,one expert from Information and Communications Technology Council (ICTC), previously told HRD that “foundational AI and technical literacy” are most effective when reinforced through “real-world problem solving that mirrors practical business cases”.
“A successful training approach starts by ensuring that leaders at all levels, including executives and managers, are AI-literate enough to identify meaningful opportunities for adoption,” said Todd Legere, one of ICTC's Research & Policy Analysts, in a statement sent to HRD.
“Beyond technical depth, organizations should foster broad AI literacy so employees understand both the potential and the limits of AI, and how to use it responsibly with appropriate human oversight, which is essential for building trust in AI systems and encouraging their responsible, effective adoption throughout the organization.
Meanwhile, CPHR Canada board chair Erin Polcyn Sailer previously told HRD that AI-skilled workers “won't be able to add value” if AI is simply an “add-on or a response to market hype”. She cautioned that "too often, organizations focus only on hiring technical specialists," adding: "Technology alone doesn't create value – people do."
