Best AI Prompts for Brainstorming Sessions with Claude

TL;DR

• Claude excels at structured evaluation and synthesis during brainstorming, making it uniquely suited to help teams move from a large idea set to clear prioritized decisions.
• The most effective Claude brainstorming prompts combine divergent generation (many ideas) with a built-in convergent framework (evaluation criteria) to prevent idea overload.
• Role-play prompts where Claude adopts a specific persona (skeptic, customer, competitor) generate more diverse perspectives than generic ideation.
• Claude’s ability to hold long conversations makes it ideal for multi-round brainstorming where each round builds on the previous one.
• Post-brainstorming synthesis prompts help distill dozens of generated ideas into thematic clusters and ranked recommendations.

The fundamental bottleneck in most brainstorming sessions is not idea generation; it is idea synthesis. Teams are reasonably good at producing long lists of ideas. They are terrible at turning those lists into decisions. Claude’s strength is not just generating ideas but holding the full context of a brainstorming conversation and helping a team see patterns, make tradeoffs, and converge on action. This guide focuses on the prompts that exploit that analytical capability, not just Claude’s generative creativity.

1. The Divergent-Convergent Prompt Structure

The most productive AI-assisted brainstorming sessions use a deliberate two-phase structure. Phase one is divergent: generate widely and do not filter. Phase two is convergent: apply rigorous evaluation to narrow to the best options. Claude can run both phases, but the prompts for each phase look very different.

Divergent phase prompt:

I am leading a 45-minute brainstorm with my product team on how to reduce our B2B SaaS churn rate, which currently sits at 8% annually. I want you to act as an idea generation facilitator. Generate 25 ideas across the following categories: product improvements that would increase value retention, lifecycle email or in-app communication changes, pricing or packaging changes, customer success intervention strategies, and self-service onboarding or education improvements. Do not filter for feasibility yet. I want the widest possible range, including ideas that might seem unconventional or challenging to implement. Label each idea as: product, communication, pricing, CS, or education.

This divergent prompt explicitly defers filtering and specifies five categories to ensure breadth. The category labels make the convergent phase easier because ideas are already organized.

Convergent phase prompt:

Now I have a list of 25 churn reduction ideas you generated. I want you to act as a strategic advisor and help me narrow this to the top 5 for a 90-day experiment roadmap. Apply the following evaluation criteria to each idea: (1) estimated impact on churn rate if it works (low/medium/high), (2) implementation cost and complexity (1-3 scale, 3 being most complex), (3) speed to test (how quickly can we run an experiment to validate?), (4) whether it addresses a root cause vs. a symptom of churn. Evaluate all 25 ideas against these criteria, then select the top 5 that represent the best combination of impact and feasibility. For each of the top 5, provide a one-sentence experiment hypothesis and suggest the smallest possible test design.

The convergent prompt adds specific evaluation dimensions that go beyond a simple impact-versus-effort matrix. Speed to test is particularly important because it prioritizes ideas that generate learning quickly, regardless of their ultimate impact.

2. The Competitor Role-Play Prompt

One of Claude’s strengths is adopting perspectives convincingly. Role-play prompts where Claude takes on a specific persona generate critiques and ideas that feel more external and diverse than the team’s own thinking.

Prompt for competitor perspective role-play:

Imagine you are a senior product strategist at our primary competitor, a well-funded startup with a comparable product and a different philosophical approach to customer success. We are a mid-market B2B SaaS company with $8M ARR and 75% NRR. Our churn rate is 8% annually and our competitor's is around 4%. I want you to generate a frank assessment from the competitor's perspective: what would they say are the three structural weaknesses in our approach to churn that allow them to win on retention? Be specific and pointed, not diplomatic. This is a competitive intelligence exercise.

Then, from the same competitor perspective, suggest one strategic move they are likely making in the next 6 months that would further widen the retention gap if we do not respond.

Competitor role-play prompts generate a specific type of external perspective that teams rarely access. They also tend to surface uncomfortable truths that internal stakeholders would soften or avoid.

3. Assumption-Surfacing Prompts

Every strategy rests on assumptions. The most dangerous ones are the ones nobody in the room articulated because they are too foundational to question. Claude can surface these systematically.

Prompt for assumption surfacing:

I am preparing to pitch a new enterprise sales motion to our board. The pitch will argue that we should shift from a product-led growth model to a field sales model over the next 18 months, targeting 10 enterprise deals at $150K+ ACV in the first year. I want you to surface the key assumptions underlying this strategy. For each assumption, I want you to identify: what evidence currently supports it, what evidence would contradict it, what the consequence is if the assumption is wrong, and whether this assumption has been validated by any external data or expert opinion. Focus on assumptions about the enterprise market, our product's readiness for enterprise requirements, our team's ability to execute enterprise sales, and customer willingness to switch from our current product-led motion.

Assumption surfacing is the most underrated pre-decision activity in strategy. It converts implicit beliefs into explicit hypotheses that can be tested or protected.

4. Six Hats Thinking with Claude

Edward de Bono’s Six Hats framework assigns different thinking modes to each “hat”: information (white), emotions (red), caution (black), optimism (yellow), creativity (green), and process control (blue). Claude can systematically run through all six hats on a given problem, producing a more balanced analysis than a single-mode human discussion.

Prompt for Six Hats analysis:

Apply de Bono's Six Hats framework to our decision about whether to build a native mobile app or continue optimizing our responsive web app for mobile browsers. We have 40% of our users accessing via mobile browsers, mobile revenue is 15% of total, and a native app would require approximately 6 months and 2 dedicated engineers to build a v1 with feature parity to web.

Go through each hat: (1) White Hat: what data and facts do we have about our mobile users and their behavior? (2) Red Hat: what gut reactions do our users likely have about a native app vs. web? (3) Black Hat: what are the most serious risks and drawbacks of building a native app? (4) Yellow Hat: what are the potential upsides and opportunities a native app could create? (5) Green Hat: what alternative creative solutions might address the core need without building a full native app? (6) Blue Hat: what is the synthesis and recommendation given all the above perspectives?

The Six Hats prompt is particularly valuable for decisions that have strong advocates on multiple sides, because it forces consideration of all perspectives rather than letting the most confident voice dominate.

5. Reverse Brainstorming Prompt

Instead of asking “how do we achieve X?” ask “what would cause us to fail at X?” This inversion often surfaces risks and obstacles that the positive framing misses.

Prompt for reverse brainstorming:

Instead of asking how to improve our net revenue retention from 85% to 95%, I want you to do reverse brainstorming. List 15 specific ways we could make NRR worse over the next 12 months. Be specific: not "we could lose customers" but "our enterprise customers could experience service degradation when their IT departments tighten security policies, leading them to freeze new licenses." For each failure mode, identify what the early warning indicators would be in our data (metrics, signals) and what the fastest recovery path would be if we saw that indicator trending negative.

Reverse brainstorming is particularly useful for risk-averse decisions or strategic bets where the downside is severe. It does not replace positive ideation but supplements it with a defensive perspective.

FAQ

How is Claude different from ChatGPT for brainstorming? Claude is generally stronger at sustained analytical reasoning across a long conversation, which makes it better for multi-round brainstorming where each prompt builds on previous outputs. It also tends to be more structured and less prone to generic responses when given specific evaluation criteria.

Should I use Claude’s brainstorming output directly in presentations? Use it as a starting point for your team’s thinking, not as finished recommendations. The value is in the perspective and the framework; the judgment about what fits your specific context should come from your team.

How do I get Claude to generate ideas in my industry’s specific language? Include industry terminology and context in your prompt. The more specific vocabulary you provide, the more accurate the generated ideas will be to your domain. Also ask Claude to use specific examples from comparable companies in your space.

What do I do if my team is attached to an idea that Claude’s evaluation says is weak? Use the adversarial challenge prompt to surface the weaknesses explicitly in front of the team. Make it a shared exercise where Claude plays devil’s advocate and the team responds with counterarguments. If the team can rebut the critique with evidence, the idea may be stronger than the evaluation suggested.

How many rounds of brainstorming should I run with Claude? Three rounds is usually the sweet spot: (1) divergent generation, (2) convergent evaluation, (3) synthesis and recommendation. Each round should be a separate prompt that builds on the previous output rather than a single mega-prompt trying to do everything.

Conclusion

Claude’s comparative advantage in brainstorming is not idea volume but structured evaluation. The tools in this guide are designed to exploit that analytical strength: generate widely, evaluate rigorously, stress-test assumptions, and synthesize into actionable recommendations.

Key Takeaways:

• Use a deliberate divergent-convergent structure to prevent idea overload from blocking decisions.
• Role-play prompts (competitor, customer, skeptic) generate external perspectives that balance internal groupthink.
• Assumption-surfacing should be a standard pre-decision activity for any significant strategic bet.
• Six Hats analysis prevents any single thinking mode from dominating a complex decision.
• Reverse brainstorming complements positive ideation by surfacing failure modes and recovery paths.

Next Step: Run the divergent-convergent prompt structure on your next open strategic question. Start with the divergent phase, then paste the generated list into the convergent phase. Compare the AI-prioritized top 5 to your team’s intuition about priorities. The delta is a productive conversation about what the team knows that the AI does not, and vice versa.