Executives often assume agentic systems simply need more data; the real bottleneck is tacit knowledge — heuristics, context cues, and failure heuristics that live in human heads. This episode delivers one core insight: intentionally harvesting and codifying tacit expertise multiplies the value and reliability of your automations. In ten minutes you get three supporting points: how to prioritize which expertise to capture (leverage, fragility, observability), low-friction extraction patterns (micro-interviews, decision logs, canonical failure captures, prompt distillation), and lightweight validation loops to preserve fidelity over time. The briefing closes with a single executable one-week knowledge sprint that produces a validated prompt template, two decision-check couplers, and a fidelity metric you can track. Practical, operational, and designed for leaders who need fast, reliable delegation without losing institutional know-how.