Prompt
You are a project planning assistant. Turn the project below into a draft work breakdown structure and high-level plan. This is a starting point for me to refine with my team — not a committed plan.

Project goal and definition of done: {{project_goal}}
Key deliverables and known scope: {{deliverables}}
Constraints (deadline, budget, team, methodology): {{constraints}}

Produce:
1. A WBS to 2-3 levels: phases -> deliverables -> tasks.
2. Suggested milestones and the major dependencies between phases.
3. A rough sequencing view (what must finish before what).
4. An "Assumptions and open questions" section.

Rules:
- Base the plan only on the scope I gave you. Do not add deliverables, roles, or work packages I did not mention; if something important seems missing, raise it as a question instead of inventing it.
- If you include any duration or effort estimate, label it "ROUGH — confirm with the team" and explain it is not based on our historical data.
- Do not present the timeline as achievable or committed — that requires the team's input and my sign-off.
- Keep task names short and outcome-focused.

Fill in your details and the prompt updates live — then copy.

What you get back (excerpt)

Phase 1 — Discovery - Map top tier-1 ticket types (deliverable: prioritized list) - Define self-service success metric with support lead Phase 2 — Build - Knowledge base (author top 20 articles) - Account login + ticket status lookup Milestones: Discovery sign-off; KB content-complete; beta with 10 users. Dependency: Login depends on IT provisioning SSO. Assumptions and open questions: - Assumes existing help content can be reused — confirm. - Estimates omitted; durations are ROUGH — confirm with the team.

The full workflow

  1. Give the prompt your real goal, deliverables, and hard constraints
  2. Use the WBS as a first draft — cut invented tasks and add the ones the model missed
  3. Size every task with the people who will do the work and against past-project actuals, not the AI's numbers
  4. Load the agreed plan into your PM tool and baseline it only after sponsor sign-off

Watch out for

AI durations and effort estimates are not grounded in your team's velocity or historical data — treating them as real is how plans go off the rails. Validate every estimate and dependency with the team, and remember scope and budget commitments are the sponsor's call, not the model's.

Keep confidential scope, client names, and commercial terms out of consumer AI tools; use generic descriptions or an employer-approved enterprise account.

Where this comes from

Every use case on this site is grounded in real reports from working project managers — not invented by us.

More AI use cases for project managers

← All 6 use cases: How Project Managers Use AI