AI Planning, Human Decisions

An AI-generated plan can make a blank page easier to approach. It can suggest milestones, break down a broad objective, and surface questions for discussion. But a detailed draft can also look more certain than the information behind it. Treating that draft as a proposal keeps the team responsible for what it commits to deliver.

Give the draft useful context

Describe the customer problem, the desired outcome, and the constraints that matter. Include team capacity, known dependencies, and what is outside the scope. Without those boundaries, a plausible plan may describe a project your team cannot realistically undertake.

Ask for assumptions to be listed separately from tasks. This makes it easier to distinguish information supplied by the team from details that were inferred to fill a gap.

Review the risky parts first

Start with estimates, sequencing, permissions, and dependencies. The people closest to the work should check whether the proposed steps are feasible. A task that sounds reasonable may still depend on access, research, or another team’s decision.

Use the draft to improve the conversation. Ask what is missing, what could be removed, and which milestone needs evidence before a date is assigned. Do not equate a longer task list with a more reliable plan.

Keep a person accountable

Give each accepted milestone a named owner. That person is responsible for maintaining the plan and raising uncertainty, even when several people contribute to delivery. AI assistance does not replace that accountability.

When the plan changes, record the reason. A reviewed decision history is more useful than repeatedly generating a fresh plan that forgets the tradeoffs the team has already made.

A practical next step

For your next planning session, bring an AI draft alongside a short list of assumptions. Judge the session by the decisions the team resolves, not the number of tasks it accepts.