Journal / Kirill Naumov
Why employees are not using AI — and what leaders should check
Distinguish skill gaps from workflow, trust and access problems when introducing AI at work.
By Kirill Naumov ·
The tool is purchased, training is completed and the team returns to old habits. That is not necessarily resistance to change. The new approach may simply be harder.
Examine the whole task
An answer may arrive quickly, only to require lengthy checking and copying into another system. Observe the end-to-end workflow, not generation alone. Remove unnecessary steps before demanding more activity.
Check trust and accountability
Do colleagues know which data they may enter and who is accountable for an incorrect answer? Is there a tested example for their role? When rules are unclear, avoiding a tool can be a sensible decision.
Practise by role
A generic webinar does not replace working through a real task. A store manager, trainer and commercial analyst need different scenarios. Start with permitted data and demonstrate an error and how to detect it, not only a successful output.
Do not turn prompt counts into a target
Mandatory prompt quotas reward activity rather than value. Look at voluntary repeat use, work quality and total task time including review. Ask what colleagues no longer need to do because the tool helps.
Assign ownership
Name a use-case owner, establish feedback and schedule a review. If value is absent, change or stop the experiment. Being able to stop a poor idea builds more trust than another innovation presentation.