AI is either a moat or a mop. Learn to classify: proprietary-data workflows where AI is durable advantage, and generic workflows where it is table stakes — and why the two deserve different budgets and different urgency.

Every AI demo is edited. Learn the audit questions — data freshness, failure rates, pricing at your scale, the exit cost — and the pilot-before-procure rule that converts demos into evidence.

Privacy, liability, and sector rules move faster than strategy decks. Build the habit: which data classes can never train a vendor model, who is liable when the AI errs, and the standing check before every launch.

Strategy is what you will NOT do. Assemble the page: three advantage bets with owners, the hygiene list with deadlines, guardrails, and the kill criteria — reviewed quarterly, rewritten yearly.