Why Only 13% of Companies Get Real AI Productivity Gains

The study exposed an uncomfortable dynamic. Workers in top-performing AI orgs are 18% more likely to refrain from using AI on certain tasks. They are also more likely to bend the rules: 54% use unapproved tools or use approved tools in noncompliant ways; 36% hide how much AI is helping them. The reason, the authors noted, is that official systems are "too slow, too narrow, or too disconnected from how the work actually gets done."

Training and rewards matter. In top AI organizations, 90% of workers said their employer treats AI as a chance to redesign work, versus 54% in lagging organizations. 90% said their employer provides enough AI training, compared with 52% at less-engaged firms. 84% said their employer formally rewards AI skills, versus 48% of the laggards.

For investors, the study clarifies which companies are likely to see a real return on AI spending. The ones that layer AI on top of broken processes will burn money and morale. The ones that redesign workflows – and invest in context, training, and the willingness to let workers break the rules when the official path is slower – are the ones that will close the gap between hours saved and performance gained.

The study's own language is blunt. "Most organizations will keep learning the hard way that AI's time savings aren't free," the authors wrote. "The hours workers save come back as botsitting. The judgment they offload comes back as botshitting." The 13% know that the hard work is not buying AI – it's changing how work works.