Huntley: protected AI practice turns adoption into a management responsibility

Geoffrey Huntley argues that managers who leave AI learning to spare time are preserving an obsolete performance baseline. His proposed mechanism is protected professional-development time, visible experimentation, and enough hands-on agent knowledge for engineers and leaders to explain the tools. Why it matters: An AI mandate without time to practice rewards employees who can learn off-hours and leaves everyone else accountable for a transition the organization has not enabled. Treating experimentation as scheduled work makes adoption a management and operating-model decision rather than a private productivity contest.

Try this: Run a two-hour team lab in paid work time: have each engineer build or modify one small tool-calling agent, then inspect the code, context flow, tool permissions, and verification evidence together. Record one process constraint the team should change before the next lab.