90% of Firms Report Zero AI Productivity Impact. Yet 80% Already Laid Off Workers.

90% of firms report zero productivity impact from AI. Yet 80% of large enterprises have already laid off workers after launching AI initiatives.

The NBER study from February found that over 90% of companies saw no measurable change in employment or output from AI adoption. Less than 30% of CEOs expressed satisfaction with their generative AI returns.

Gartner published data that makes it worse: enterprises with significant AI ROI laid off workers at the same rate as enterprises with negative AI ROI. There is no correlation between AI returns and workforce reductions. None.

The companies getting the most ROI from AI are not the ones cutting headcount. They are the ones reinvesting in their workforce — training employees to build their own agents, create their own automations, and orchestrate AI systems.

As one CTO put it: "You cannot automate expertise you no longer have."

Three patterns in enterprises that fail at AI deployment: they measure progress by headcount reduction instead of capability expansion, they automate processes before understanding whether those processes should exist, and they treat AI adoption as a cost-cutting exercise instead of a reinvestment decision.

Layoffs are being used as a proxy for AI progress. They are not progress.