A good answer solves today's problem. A good way of thinking solves the next twenty. AI-native founders move fast and face problems no one has a playbook for yet, so the most valuable thing you can give them is the reasoning they'll use when you're not in the room.
This means narrating how you think: what you look at first, which numbers you trust, how you weigh a reversible decision against an irreversible one, when you'd stop analysing and act. Founders learn as much from how an experienced operator reasons through a trade-off as from the conclusion.
Measure your mentorship by how quickly the founder stops needing you for a given type of problem. If they're coming back with the same question three months later, something went wrong.