How to Brainstorm Ideas for a Project with Copilot
Copilot can generate 12 ideas faster than you can generate 3. Use it to break out of stuck thinking and pressure-test what you have.
What it is
Brainstorming with Copilot isn’t a replacement for human creativity — it’s a prompt for it. Ask for 10 ideas on a topic and Copilot produces 10. Some are obvious, some are bad, a few are interesting. The interesting ones spark your own thinking. The obvious ones confirm what you already know. The bad ones, by contrast, sharpen what makes a good idea.
The skill is iterating. First-draft brainstorming is broad and generic. Pick the most interesting 2-3 ideas and ask Copilot to expand them: ‘For idea 3, what would the first month of work look like?’ or ‘For idea 7, what could go wrong and how would I handle it?’. The deeper questions produce more useful output than the initial brainstorm.
Copilot is also good at the meta question: ‘What am I missing?’ or ‘What’s a non-obvious approach to this?’. These prompts get past the boring options and surface unconventional thinking. They won’t always work, but when they do, they save you from going down the same well-worn path everyone else takes.
When to use this
- When you’re stuck on a problem and need fresh angles.
- When you need a list of options to choose from.
- When you want to pressure-test an idea you’ve already had.
- When you’re starting a project and need to generate possibilities before narrowing down.
How to do it
- Explain the goal, constraints, and what ‘good’ looks like.
- Ask Copilot for 10-12 ideas — quantity beats quality at this stage.
- Ask Copilot to group the ideas into themes.
- Pick the top 2-3 and ask for pros/cons, effort, and risks for each.
- Ask ‘what am I missing?’ or ‘what’s a non-obvious option?’.
- Pick one idea and ask for a simple plan to execute.
- Apply your own judgement to the final selection — Copilot generates options, you choose.
Best practices
- Brainstorm wide first, then narrow. Don’t kill ideas in the first round.
- Ask ‘what am I missing?’. Reveals blind spots in your thinking.
- Iterate on the top 2-3 ideas. Depth beats breadth after the initial brainstorm.
- Capture ideas somewhere durable. Otherwise they evaporate when the chat ends.
Common mistakes
- Stopping at the first round. Initial brainstorming is generic. Push for round two and three.
- Letting Copilot pick for you. Generation is the AI’s job; selection is yours.
- Brainstorming without constraints. ‘What should we do?’ is unanswerable. Add constraints (time, budget, scope) for sharper output.
The Copilot Readiness Guide gives you the 25-question scorecard, the 4-category risk audit, and the 30-day plan to fix permissions, content quality, and sensitive content before go-live.
Get the Copilot Readiness Guide — $39 →FAQ
Is Copilot good for brainstorming?
Yes — for generating quantity quickly. Less good at originality (most ideas are recombinations of common patterns) and at challenging your assumptions. Use Copilot to expand your idea space rapidly, then filter ruthlessly. Don’t outsource taste.
What’s the best Copilot prompt for brainstorming?
Specific and constrained: ‘Give me 10 marketing angles for a product launch targeting busy parents who don’t have time to research. Avoid the obvious ones like time-saving — push deeper.’ Vague prompts (‘give me ideas’) get generic output; constrained prompts force Copilot past obvious answers.
How do I get more original ideas from Copilot?
Tell it what to avoid: ‘No clichés about [topic]. No mentions of [common ideas].’ Give examples of bad ideas to rule out the obvious answers. Ask for ideas ‘from unusual angles’ or ‘that would surprise the team’. Copilot defaults to safe; you have to push it to surprising.
Should I use Copilot for creative or critical brainstorming?
Creative — yes, with caveats. Critical brainstorming (questioning assumptions, finding what’s wrong) — less effective. Copilot tends to be agreeable; if you need devil’s advocate, give it explicit instructions (‘argue why this idea will fail’) or use a different model with more pushback baked in.