Evaluating and Screening AI Skills Effectively for GTM Hiring Beyond Self-Reported Proficiency

curious how people here are evaluating AI skills when hiring for GTM roles. everyone lists 'AI proficiency' as a requirement now but when we actually assess people's AI skills conversationally, most score way lower than their self-reported level. we ran 412 assessments and the average was 52/100 -- 62% scored below Proficient. anyone found a reliable way to screen for actual AI competence beyond 'tell me about a time you used AI'?