How Do GTM Teams Effectively Assess Real AI Proficiency Beyond Self-Reporting?

genuine question for the GTM folks here — how are you evaluating whether your team actually knows how to use AI tools vs. just saying they do? we've been seeing a massive gap between self-reported AI proficiency and tested skill levels, especially around knowing when AI output is wrong. curious if anyone's built this into their hiring or enablement process, or if it's still vibes-based