1. Take AI off the pedestal first.
Make the limits visible before demonstrating the power. People use tools better when they understand where the blade is.
Never pitch. Only teach.
A field-tested method for introducing technology to people who did not ask for it and have legitimate reasons not to trust it. The method was shaped in rail operations, safety-critical IoT deployments, and regulated enterprise work.
Resistance is often treated as a people problem. Usually it is a signal: the rollout has not answered the employee’s real fear, preserved their control, or demonstrated value in their work.
Make the limits visible before demonstrating the power. People use tools better when they understand where the blade is.
“I do not trust it” may mean this makes me look stupid, adds work, removes control, or threatens my job.
Put the employee in the driver’s seat before asking for behavior change. Their judgment should become more important, not less.
Useful learning creates its own proof. A skeptical room can smell a pitch immediately.
AI can clear repetitive work and still make mistakes. Treat the output like a confident new hire’s report: useful, fast, and requiring review.
Do not defeat the person the room follows. Give them the role the system cannot perform: judgment.
I spent years as the operator expected to use technology designed by someone else. Later I built and delivered adoption programs for mechanical teams, maintenance, supervisors and executives, and enterprise delivery organizations. I do not ask employees to trust a tool blindly. I teach them how to use it without surrendering control.
Identify what the team actually does, where AI is entering the process, and what employees are afraid of losing.
Use examples that resemble the team’s real decisions rather than generic prompt demonstrations.
Leave behind a simple tool, checklist, or review standard that reinforces the behavior after the workshop.
Recruit credible practitioners, including skeptics, to model the standard and improve it from the field.
The explanation, process, ownership, or handoff usually failed first. Good adoption begins by fixing those things.