McKinney Surveyed 1,719 Executives. 94 Percent Learned Nothing. Spending Without Systems Is Just Spending.

McKinsey's 2026 survey of 1,719 executives reveals that 94% of enterprises cannot demonstrate significant earnings impact from AI despite record spending. The gap is telling: 80% of individual workers report productivity gains, but only 6% of organizations translate that into bottom-line results. The distinguishing factor among top performers is agentic AI deployment, with 40% of organizations earning above $1 billion annually now scaling agents in at least one function, up from 27% the prior year. Smaller organizations remained flat at 22%.

This is what organizational theorists call the 'diffusion bottleneck.' Individual productivity gains do not automatically compound into enterprise value. The mechanism is structural: without rearchitecting workflows around autonomous agents, you are simply giving workers a faster calculator and wondering why the spreadsheet did not improve. Well, actually, the spreadsheet did improve. You just did not change what happens to the spreadsheet.

McKinsey conducted the survey of 1,719 executives. Two-thirds of respondents reported little or no AI-related change in total employment, which should temper your enthusiasm about both mass layoffs and mass gains.