Brainstorming with ChatGPT

ChatGPT can be used as a structured thinking partner that's especially useful when starting from a blank page, organizing complex options, or quickly producing a draft. Rather than being a tool that judges ideas for you, it's more of an assistant that speeds up your thinking and helps organize direction.

Brainstorming tends to get stuck at two points: when you have too few ideas, or conversely when you have too many but no structure to organize them. ChatGPT is good at the following three things.

• Expanding options: quickly generating many angles, experiments, messages, and alternatives.
• Adding structure: grouping ideas by theme, suggesting simple frameworks, and turning vague goals into selectable options.
• Pressure testing: surfacing a plan's assumptions and trade-offs upfront, so you can check weaknesses before actual execution.

An effective approach is the wide → narrow flow. First, generate ideas broadly by attaching constraints, then group and compare them by theme, and finally narrow down to an execution plan. When starting out, it's best to clarify "what needs to be decided" and provide constraints such as audience, timeline, capacity, and success criteria, as well as any previous attempts.

Here are also some prompt manipulations that improve the quality of thinking.

• Asking for reasons: have it explain why it recommends a particular option.
• Having it pick one: ask which one it would choose if it could only pick one.
• Requesting friendly criticism: have it point out what could be made stronger.
• Separating quick wins from foundational work: distinguish what can be done right away from foundational work.
• Using evaluation criteria: score options using criteria like impact, effort, and confidence.
• Changing format: restructure into a 2x2, decision tree, timeline, or stakeholder map.
• Using voice input: input rough, unorganized thoughts and have it organize them into themes and next steps.

Examples of tasks this can be applied to directly include team offsite session ideas, product launch campaign themes, internal process improvements, improving cross-team collaboration, uncovering missed opportunities, next-quarter planning, and creating paths for high-stakes decisions. The key is to separate idea generation from evaluation, and to treat ChatGPT's output not as a final answer but as a reviewable draft.