29 - AI PM: trust & evaluation | Shira Weinberg Harel

build trust with users and stakeholders - transparency and explainability, reliability and consistency (how to fail safely), automation vs user control. how to evaluate with uncertainty, prompt engineering, effective bug reports