Brainstorm and plan a project
Go from a vague idea to a clear plan with steps, owners and deadlines.
- Brainstorm: "Give me 15 ideas for [goal]. Mix safe and bold ones."
- Narrow down: "Pick the best three for a small team with £2,000 and explain why."
- Plan: "Make a step-by-step plan working back from [date], with deadlines."
- Spot risks: "What could go wrong, and how do we prevent it?"
- Make it usable: "Turn this into a checklist" or "a table I can paste into a spreadsheet".
Share it with your team and adjust: AI gives you a strong first version, not the final word.
Brainstorming that's actually useful
- Quantity first: "Give me 20 ideas" gives more range than "give me an idea".
- Mix it up: "Include some unusual ideas" or "ideas a competitor wouldn't try".
- Use constraints: budget, time, team size and audience make ideas realistic.
- Combine: "Combine ideas 3 and 7 into something stronger."
From idea to plan
- Define success: "What would make this project a success in three months?"
- Ask it to ask you questions before planning; it will spot gaps.
- Work backwards from the deadline: phases, tasks, dependencies, owners.
- Format: a table (task, owner, start, due, status) pastes neatly into a spreadsheet.
- Risks and assumptions: list them explicitly.
- Communication: "Draft a short kick-off email explaining the plan to the team."
Keeping it on track
Each week, paste your progress and ask: "What's at risk of slipping, and what should I focus on this week?"
Thinking tools AI can run for you
- Pre-mortem: "Imagine this project failed six months from now. Write the most likely reasons." Then plan to prevent them.
- Options appraisal: "Compare these three approaches on cost, time, risk and impact in a table."
- Stakeholder map: "Who is affected, who decides, who could block this, and what does each need?"
- Assumption testing: "Which assumptions in this plan, if wrong, would cause the biggest problems? How can we test them cheaply?"
Better plans
- Dependencies: ask it to flag tasks that can't start until others finish, and the critical path (the chain that sets the end date).
- Buffers: "Add realistic contingency for approvals and supplier delays."
- Capacity: give holidays and part-time hours so timelines are realistic.
- Milestones: a few clear checkpoints are easier to track than dozens of tasks.
Tools
Your assistant can produce tables for spreadsheets, but project tools and some AI agents can create tasks directly in planning apps. Keep a human reviewing the plan; AI doesn't know your organisation's politics, approval routes or unwritten rules.
Pitfalls
- False precision: neat timelines can hide guesswork. Mark estimates as estimates.
- Generic plans: insist on your specifics (team names, real dates, actual budget).
- Skipping people: a plan nobody discussed rarely survives. Use the AI draft to start the conversation.
Try this prompt
I need to [project goal] by [date]. Resources: [people, budget, time]. Constraints: [anything fixed]. 1) Ask me any questions you need first. 2) Then give me a plan with phases, tasks, who does what and deadlines, working back from the date. 3) List the top five risks and how to reduce them.
Replace the [brackets] with your own details