Why might employees keep their AI productivity improvements to themselves?
Short answer: Because sharing can cost them more than it earns. A faster process can reflect effort nobody sees: time spent learning, testing, getting things wrong, and applying existing skills and judgment. If sharing it brings a heavier workload, an expectation to teach everyone else and little credit, the employee may feel their investment has been taken rather than valued. They may also worry about losing their hard-earned advantage, control over their time, or how necessary their role will seem.
Economists describe a related incentive problem as the "ratchet effect": strong performance can raise future expectations, giving people a reason to hold back. Ethan Mollick's "secret cyborgs" describes people who use AI at work without revealing it. These ideas help explain why asking for transparency is not enough.
What should leaders do?
Agree on outcomes and quality, rather than treating every saved hour as space for another task.
Explain how gains will be shared: benefits for the person, capacity for the organisation, and time reinvested in learning.
Reward the discovery, not only the extra output. Keep the discoverer's contribution visible when the process is adopted.
Address concerns about roles and headcount directly. Promise only what the organisation can keep.
The test: if your best performer showed that a task could take half the time, would they feel proud or exposed?
Where enablement fits
At Fluence AI, enablement means building confidence on the skills people already have, not simply introducing another tool. Addressing what people fear losing is part of helping them feel safe enough to contribute.
Next step: write to tshibangu@fluenceai.co.za. The first conversation is free.
More questions
Written by Tshibangu Kalambaie, Founder, Fluence AI.