Reflection — 141 AI solutions versus 3, at comparable maturity: the difference is persistence

A two-year field study followed two organizations determined to put AI to work, with comparable maturity, the same tools, and equally sensible use cases. One ends up with 141 AI solutions actually in use, the other with only three. What sets them apart isn’t the technology or access to tools, but the support given over time to the people doing the experimenting. Where AI failed to take hold, over 80% of the people who had started using it eventually disengaged, for lack of ongoing support and because their AI work stayed invisible in performance reviews.

The lesson for leaders: AI doesn’t stall for lack of adoption, it stalls for lack of persistence. Running a training session and handing out access isn’t enough; without a framework that supports and rewards the effort, the early momentum fades within weeks. Name someone to bring teams together each week around the use cases that work, and make the time spent on AI count in goals and reviews, so that effort is seen rather than quietly set aside.

Source: MIT Sloan, “Why some organizations turn AI experiments into business value while others quietly fail,” September 9, 2026 (field study, two organizations).