A Monday morning earnings call ends, and the chief financial officer opens a chat window to draft the board update in minutes. Down the hall, a frontline finance team still works the old way, because access, training and incentives for the use of artificial intelligence never arrived there.
That split reality sits at the heart of a new internationally representative firm survey of almost 6,000 CFOs, CEOs and senior executives across the U.S., United Kingdom, Germany and Australia, published in the National Bureau of Economic Research. The paper, titled Firm Data on AI, paints a picture of fast diffusion but limited realized impact, with large expected gains coming soon. Executives forecast a smaller workforce, driven largely by slower hiring rather than layoffs. But the real story for leaders lies in the gap between ambition and daily use, and the widening disconnect between executive expectations and employee beliefs.
Productivity Gains on the Horizon
Senior leaders in the survey expect AI to move the productivity needle in a way that dwarfs most operational initiatives. Across the four regions, executives forecast about 1.4 percent higher productivity over the next three years from AI adoption, with the U.S. at about 2.3 percent over the same time horizon—a pace that translates to roughly 0.77 percentage points per year. This would nearly double the anticipated baseline growth of about 1 percent annually.
Yet the paper also reports that realized impact over the past three years stayed modest, with an average realized productivity gain around 0.29 percent across firms. That “quiet period” matters, because it explains why many organizations still treat AI as a pilot program rather than a system in full operation.
Adoption Accelerates in 2025
Executives forecast acceleration because deployment patterns shifted sharply during 2025. There was a jump in usage frequency and a drop in the share reporting zero use, to roughly one-fourth of respondents. The adoption signal is clear in the paper’s executive use measures, and it sets expectations that teams will soon face new usage standards. For instance, Accenture now tracks how often senior employees utilize artificial intelligence on a weekly basis, according to recent reports. The firm links these adoption metrics to promotion opportunities for veteran staff, to ensure they embrace the growing role of technology in the workplace.
Employment Forecasts: A Slow Squeeze
Employment expectations in the survey land even harder than the productivity forecasts. Executives predict a net employment decline of about 0.7 percent over the next three years, driven largely by reduced hiring rather than mass layoffs—a point developed in employment forecasts. The authors translate that share into about 2 million jobs across the four regions. Hiring plans move early because they sit inside annual budgeting, requisition controls and backfill decisions. That means workforce change can arrive as a series of “roles never refilled” rather than a single dramatic event.
Leaders can turn that dynamic into an advantage through intentional design. Instead of letting hiring slowdowns spread as silent austerity, organizations can define where AI substitutes for throughput, where it improves decision quality, and where it frees capacity for growth.
Macro Context and Inequality Risks
Macro research supports that framing. The International Monetary Fund estimates that a large share of jobs are in roles touched by AI, and it emphasizes that productivity gains can arrive alongside inequality risks. The International Labor Organization’s global assessment of GenAI and jobs points to substantial exposure in clerical work and a strong likelihood of task reshaping, reinforcing the case for role-by-role redesign rather than blanket job narratives.
The Perception Gap
The survey’s most revealing finding may be the perception gap between executives and employees. Employees surveyed separately predict far smaller impacts, and in some cases expect employment to rise, whereas executives forecast contraction—a divergence captured in the paper’s worker-executive comparison. That gap signals an execution challenge, because scaling AI requires shared beliefs about what “good work” looks like under new tools.
External surveys show similar tension. The Organization for Economic Cooperation and Development’s work on AI and job quality emphasizes uneven adoption and the importance of worker involvement in deployment choices. And the World Economic Forum projects major churn alongside net gains in some categories, captured in its jobs report summary, which reinforces the need for credible internal narratives tied to specific roles.
For professionals, the practical move involves measuring reality faster than competitors do. Instrument adoption by role and workflow, link usage to cycle time and quality outcomes, and treat spending on AI as a capital allocation choice with clear unit economics. Pair that rigor with a talent agenda that preserves trust, since reduced hiring can make it feel like the ladder is being pulled up. Public estimates of automation exposure can help leaders calibrate scenarios, including a widely cited projection on job-task exposure that many boards already reference.
The executive survey offers a clear message: Expectations already run high, and habits still lag. Organizations that close that gap first will capture the gains their leaders forecast, and they will do it with a workforce that understands how to grow inside the new reality.
