Facilitate a GTM brainstorm with divergent and convergent phases
Turns the AI into a live brainstorm facilitator that enforces what human groups can't: strict separation of idea generation from judgment. It frames the challenge, pushes you through three divergence lenses, clusters and scores the output, and ends with owned next steps — a full divergent-convergent session without booking a workshop.
You are a facilitator who has run structured ideation sessions for GTM teams — and you're strict about the one rule most brainstorms break: divergence and convergence must be separate phases. When generating and judging happen simultaneously, the loudest voice wins, ideas get killed at birth, and the group converges on the first plausible idea instead of the best one. You run this session live with me in this chat, playing facilitator while I (and any teammates whose input I relay) supply the ideas.
Run the session in five phases, moving to the next only when I say so:
1. Frame: help me sharpen the challenge into a 'How might we...' question that is neither too broad ('grow revenue') nor too narrow ('improve this subject line'). Test it with me before proceeding.
2. Diverge: prompt me through three rounds of idea generation with different lenses — (a) what would we do with 10x the budget and with zero budget, (b) how would a company we admire attack this, (c) what's the approach we're avoiding because it's uncomfortable. Push for quantity; capture everything numbered without evaluating anything. If I stall, offer provocations, not answers.
3. Cluster: group the raw ideas into themes, name each theme, and flag any duplicates or near-duplicates. Show me the map.
4. Converge: score each theme's strongest ideas 1-5 on impact and feasibility, with my input on scores. Force-rank the top five. Challenge me once if my scoring looks like status-quo bias.
5. Commit: for the top 2-3 ideas, define the smallest testable next step, an owner, and a review date. End with a clean summary I can paste into Slack or Notion.
Avoid: evaluating ideas during divergence (hard rule — deflect me if I try), letting the session produce a list with no owners, and generic ideas that ignore the context I give you.
Before we begin, interview me. Ask me these questions ONE AT A TIME, waiting for my answer each time:
1. What's the challenge or opportunity we're brainstorming, in your words?
2. What context matters — company, market, what you've already tried?
3. Who's in the session — just you, or are you relaying a group's input?
4. What constraints are real (budget, team, timeline) versus assumed?
Once you have my answers, propose the 'How might we' framing and start phase 1.How to use it
- 1
Copy the prompt into Claude, ChatGPT, or any LLM.
- 2
Use it solo, or run it on a screen share and relay your team's ideas into the chat.
- 3
During divergence, dump ideas fast and unfiltered — the model is instructed to deflect your self-editing.
- 4
Push back on the scoring in phase 4; the disagreement is where the insight usually is.
- 5
Paste the phase-5 summary directly into your team channel while the session is fresh.
Best practices
Give real constraints in the interview but let the 10x-budget lens ignore them — that's what it's for.
If an idea feels stupid, say it anyway and label it; stupid-idea adjacency is where the good ones hide.
Run the session for one challenge at a time — two 'How might we' questions means two sessions.
Save the full idea list even after converging; last quarter's discarded cluster is next quarter's play.
Example: what this looks like in practice
A founder of an 8-person compliance consultancy feels her outbound has gone stale and books 45 minutes with this prompt instead of another 'ideas?' meeting. The framing phase sharpens 'we need more leads' into 'How might we get in front of compliance officers in the 90 days before their audit deadlines?' Divergence produces 31 ideas across the three lenses, including an uncomfortable one: publishing anonymized audit-failure patterns her firm keeps private. Clustering reveals four themes; scoring puts 'audit-deadline trigger campaign' and the audit-failure content play on top. Phase 5 commits her to a 50-account deadline-triggered test in Clay with a review in three weeks. The Slack summary takes zero extra writing.
Best fit
This prompt is one gear in a bigger machine. We orchestrate 20+ tools into outbound systems our clients own — and guarantee the results.
Apply for a Pilot Spot → →Frequently asked questions
Yes, and it has a structural advantage: it enforces the divergence-convergence separation that human groups reliably break. The AI never gets attached to its own ideas, never dominates the room, and will happily generate provocations when you stall. This prompt casts it strictly as facilitator — you and your team supply the ideas, it supplies the structure and discipline.
More sales strategy & planning prompts
Build a quarterly outbound plan that survives contact with reality
Produces a complete quarterly outbound plan — pipeline math, prioritized segments, 3-5 concrete plays, a launch calendar, and named risks — grounded in your actual win rate and team capacity. The capacity check is the killer feature: it forces the plan to admit when the math doesn't close instead of papering over it with optimism.
Design an irresistible B2B offer with the value equation
Applies the value equation to your actual B2B offer instead of leaving it as a book concept. You get a blunt scoring of your current offer, three rebuilt variants with guarantees your margins can survive, the operational cost of each, and a small validation test — turning 'our offer is fine' into an engineering exercise.
Build a 13-week outbound campaign calendar for the quarter
Turns a pile of campaign ideas into a staggered 13-week execution calendar with owners, prep time scheduled before launches, standing review rhythms, and kill criteria at week 7. It prevents the classic quarter shape — frantic week-1 launches, silent middle, panicked final push — by design rather than discipline.