I need to lead a brainstorming session about [TOPIC] for [PARTICIPANTS]. The session must support this decision or outcome: [DECISION TO MAKE]. Context: - What has already happened or been tried: [CURRENT CONTEXT] - Non-negotiable constraints, including budget, timing, scope, policy, or technical limits: [CONSTRAINTS] - The priorities we should optimize for: [PRIORITIES] Create 25 brainstorming questions, not answers. Group them into five sections: clarify the problem, understand people affected, generate approaches, test assumptions and risks, and choose next experiments. Within each section, sequence questions from broad to specific. Make every question open enough to produce several possible answers, but anchored to the context above. Include questions that distinguish symptoms from root causes, identify who gains or loses, surface trade-offs, challenge the current framing, and turn promising ideas into small testable experiments. Do not use filler such as “What can we do better?” or questions whose answer is already supplied in my context. Then add a short facilitation plan: the best opening question, three questions to use if discussion stalls, and the final three questions that convert ideas into owners and next steps. Keep the full output under 900 words. Before answering, check that each question could change a decision, assumption, or experiment. Also flag any question that depends on facts not supplied here rather than silently assuming them. Ask up to three clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [TOPIC] | The issue, opportunity, or project the group needs to explore. | Reducing customer churn in the first 60 days after signup |
| [PARTICIPANTS] | Who will join the session and their relevant roles. | Head of customer success, product manager, lifecycle marketer, and support lead |
| [DECISION TO MAKE] | The decision, direction, or outcome the brainstorming should inform. | Which two retention experiments to run next quarter |
| [CURRENT CONTEXT] | Known facts, prior attempts, and the reason this needs attention now. | Churn is 18% in the first two months; new users cite setup difficulty and unclear value; onboarding emails were revised last quarter |
| [CONSTRAINTS] | Limits the group cannot ignore, such as budget, deadline, policy, or capacity. | Six-week test window, no new engineering headcount, and no discounting |
| [PRIORITIES] | The outcomes or criteria that matter most when judging ideas. | Retention impact, speed to learn, and effort from the support team |
How to use
- Paste the prompt and replace every bracketed input with your session details.
- Read the questions before the meeting and remove any that answer themselves or duplicate existing research.
- Use one section at a time, recording ideas separately from evidence and open assumptions.
- Follow up with: “Turn the strongest ideas into a one-page experiment backlog with hypothesis, owner, metric, and decision date.”
Variations
Customer interviews
Use this when you need discovery questions for one-on-one customer conversations.
I am interviewing [CUSTOMER TYPE] about [TOPIC] to decide [DECISION]. Their likely context is [CONTEXT], and these areas are off limits or sensitive: [BOUNDARIES]. Write 15 neutral, open-ended interview questions in a natural order: warm-up, current behavior, recent concrete example, pain and workarounds, alternatives, and closing. Include probes after the five most important questions. Avoid leading wording, feature pitches, and hypothetical questions when a past-behavior question will work. Flag assumptions I should validate.
Executive workshop
Use this when senior leaders need questions that lead to a strategic choice.
Design questions for a [LENGTH] executive workshop on [STRATEGIC TOPIC]. Attendees are [ATTENDEES], and the decision needed is [DECISION]. Known constraints are [CONSTRAINTS]. Produce 12 questions in agenda order, each with its purpose and the decision it informs. Include questions on market evidence, opportunity cost, risk appetite, resource implications, and dissenting views. End with four questions that make the decision criteria explicit. Do not invent market facts or assume agreement.
Root-cause session
Use this when a recurring problem needs diagnosis before solutions.
Help a cross-functional team investigate [PROBLEM]. The observable symptoms are [SYMPTOMS], the affected process is [PROCESS], and available evidence is [EVIDENCE]. Write 20 diagnostic questions grouped by timeline, variation, handoffs, controls, and contributing conditions. Separate questions that establish facts from questions that test possible causes. Include questions that look for counterexamples and recent changes. Do not propose fixes yet or imply causation from correlation. End with the five facts we must collect first.
Tips
- Start with the decision the group must make; otherwise brainstorming questions tend to create an unranked list of interesting topics.
- Separate evidence questions from idea questions so unverified claims do not become premises for the rest of the session.
- Ask about a recent concrete event when exploring user behavior; it produces better detail than asking what people generally want.
- Reserve the final minutes for experiment questions, because a strong discussion still fails if no one owns the next test.
FAQ
How many brainstorming questions should I use in one session?
Prepare more than you expect to use, but discuss roughly 8 to 15 in a 60-minute session. Follow useful threads instead of treating the list as a script.
Can AI write questions for a sensitive team problem?
Yes, but remove identifying details and set boundaries in the prompt. Keep questions focused on systems and observable behavior rather than assigning blame.
Should I ask AI to rank the questions?
Ask it to sequence them by purpose, then have participants choose based on the decision and their available evidence.