I Let an AI Plan My Quarter. It Was Right About the Wrong Things.
An agent laid out my quarter in an hour. The plan was flawless and useless, and the gap taught me what planning actually is.
I handed an AI agent my messy notes and goals and asked for a quarter plan. An hour later I had a gorgeous timeline with milestones, owners, and dependencies. It was the most organized my year had ever looked. It was also almost entirely wrong about what I should do first.
The agent planned for the world as I described it, not the world as it is. It didn't know which client was about to churn, which task I secretly dread, or which 'priority' was someone else's urgency wearing my name. Those are the things that decide a quarter, and they live in context no prompt captured.
What it was genuinely good for was the structure: turning a pile of intentions into a format I could react to. The plan was a draft, not a decision. Once I treated it as 'here's a shape, now let me argue with it,' it became useful.
I kept the timeline and threw out half the priorities, replacing them with the three things I'd be embarrassed to let slide. The AI planned my quarter; I still had to choose it. That division of labor is the whole job now.