Learn to create prompts that solve your specific PM challenges
This is where you go from using existing prompts to building your own. These meta-prompts teach you to create custom AI tools for your specific industry, team structure, and strategic challenges.
Every PM situation is unique. Generic prompts can't capture:
- Your industry's specific constraints and opportunities
- Your organization's decision-making processes
- Your team's collaboration patterns
- Your customers' unique behavioral patterns
This collection teaches you:
- Universal prompt patterns - The structural approaches that work across domains
- Question design - How to craft conversations that unlock strategic insights
- Framework integration - Weaving proven methodologies into AI interactions
- Teaching while doing - Creating prompts that educate as they execute
By learning to build prompts, you can solve problems that don't have existing solutions yet.
Every effective PM prompt follows similar structural patterns:
- Context setting - How to frame the AI's role and expertise
- Progressive questioning - One question at a time, building context systematically
- Framework integration - How PM methodologies guide AI reasoning
- Output structuring - Ensuring professional, actionable results
- Quality validation - Built-in checks and refinement loops
| Generator | Creates | Focus Area |
|---|---|---|
| artifact-first-context-intake.md | Canonical context-ingestion method | Artifact-first context collection and gap handling |
| persona-first-decision-facilitation-loop.md | Canonical facilitation pattern | Multi-turn, persona-first decision architecture |
| a-generative-AI-prompt-builder-for-product-professionals.md | Universal PM prompts | Core prompt architecture principles |
| user-story-prompt-generator-prompt.md | Custom user story workflows | Requirements gathering methodology |
| market-requirements-generator-prompt.md | Market analysis prompts | Strategic research frameworks |
| positioning-statement-prompt-generator.md | Positioning statement prompts | Geoffrey Moore-style messaging decisions |
| proto-persona-prompt-generator.md | Proto-persona prompts | Persona hypothesis development + validation |
| customer-journey-mapping-prompt-generator.md | Journey map prompts | Scope and decision-ready journey synthesis |
| daci-chart-prompt-generator.md | DACI chart prompts | Decision ownership and governance clarity |
| storyboarding-prompt-generator-prompt.md | Visual narrative tools | Communication and storytelling |
| tam-sam-som-prompt-generator.md | Market sizing prompts | Quantitative analysis frameworks |
| stakeholder-map-prompt-generator.md | Stakeholder map prompts | Power x Interest grids and engagement planning |
| prioritization-framework-prompt-generator.md | Framework choice + scoring prompts | RICE, ICE, Kano, cost of delay matching |
| premortem-prompt-generator.md | Premortem prompts | Gary Klein prospective hindsight |
| discovery-interview-prompt-generator.md | Interview guide prompts | Mom Test question design |
| research-agent-prompt-generator.md | Custom autonomous investigation prompts | Investigation mode contract: budgets, gates, evidence labels, delta rules |
Goal: Learn how prompts are actually constructed
- Start with: a-generative-AI-prompt-builder-for-product-professionals.md
- Study the structure: Notice how it asks you questions to understand YOUR context
- Run the generator: Create a prompt for a real problem you're facing
- Analyze the output: See how your inputs shaped the generated prompt
Key Learning: Effective prompts are built through systematic context gathering.
Goal: Create specialized tools for specific PM functions
- Pick your domain: User stories, market research, competitive analysis, etc.
- Use relevant generator: user-story-prompt-generator-prompt.md or market-requirements-generator-prompt.md
- Customize for context: Add your industry, team structure, and constraints
- Test and iterate: Refine based on real-world usage
Key Learning: Specialization creates dramatically better results than generic approaches.
Goal: Weave established PM methodologies into AI conversations
- Choose a PM framework: Jobs-to-be-Done, Design Thinking, Lean Startup, etc.
- Study existing examples: See how jobs-to-be-done customer circle.md integrates Osterwalder's canvas
- Build your integration: Create prompts that guide AI through framework steps
- Add pedagogical comments: Help others understand WHY the framework works
Key Learning: Structured frameworks make AI conversations exponentially more valuable.
Goal: Build generators that create other generators
- Analyze patterns: What makes some generators more effective than others?
- Abstract the principles: Extract universal patterns from successful examples
- Create meta-generators: Build tools that help others build tools
- Contribute back: Share your meta-architectures with the community
Key Learning: Teaching others to build is the highest form of mastery.
Great prompts are built on great questions. Learn to craft questions that:
🎯 Gather Strategic Context
- "What problem are we solving and for whom?"
- "What's our business context and constraints?"
- "What does success look like?"
🔍 Uncover Assumptions
- "What are we assuming about our users?"
- "What could invalidate our approach?"
- "What don't we know that we need to know?"
📊 Drive Structured Analysis
- "What frameworks should guide our thinking?"
- "What alternatives should we consider?"
- "How do we prioritize competing options?"
Build complexity gradually:
- Broad context - What domain, what problem?
- Specific situation - What constraints, what stakeholders?
- Detailed requirements - What format, what timeline?
- Quality criteria - What makes this successful?
Design interactions that teach while they execute:
- One question at a time - Prevents cognitive overload
- Context validation - "Based on what you've told me..."
- Option presentation - "Here are three approaches..."
- Refinement loops - "Would you like to adjust anything?"
This facilitation style is designed to feel like expert guidance, not a form or interrogation. Instead of dumping advice or forcing rigid templates, run a focused adaptive conversation: one clear question per turn, then context-aware recommendations when a real decision is required. This creates better PM outcomes because users get clarity without overload, structure without bureaucracy, and momentum without losing nuance.
How to position the differentiation:
- Most tools are either unstructured ("tell me more...") or over-structured (long forms and generic checklists).
- Our approach sits in the middle: conversational enough to stay human, programmatic enough to stay reliable.
- Every turn has a job, every recommendation is tied to context, and every session closes with concrete decisions and next actions.
Interaction contract for generator prompts:
- Set expectations first: goal, time, and what will happen.
- Ask one targeted question and listen.
- At decision points, offer exactly 3 context-aware recommendations with one recommended first.
- Accept
1,2,3,1 and 3, or a custom direction. - Adapt immediately, show progress, and ask the next best question.
- Close with a concise summary: what was decided, why, next actions, and assumptions to validate.
Workload inversion rule:
- Ask for minimum viable context first.
- Do not ask users to define the full structure when the assistant can propose it.
- At early forks, propose 3 candidate scopes and let users choose.
Persona-first recommendation rule:
- Phrase each option in user/persona language first.
- Add business translation second when useful for PM decision quality.
Reusable language:
"Based on what you shared, here are the three best paths. I recommend option 1 because it gets you evidence fastest with your current constraints. Reply with
1,2,3,1 and 3, or tell me your own path, and I'll adapt."
Always test your generated prompts across:
- ChatGPT (GPT-4, GPT-3.5)
- Claude (Opus, Sonnet)
- Gemini (Pro, Advanced)
- Others as available
- Use with actual PM challenges - Not hypothetical scenarios
- Iterate based on outcomes - What worked? What didn't?
- Test with colleagues - Does it work for others in similar roles?
- Document learnings - What patterns emerged?
Rate your generated prompts on:
- Clarity - Are instructions unambiguous?
- Completeness - Does it gather sufficient context?
- Usability - Can non-experts use it effectively?
- Learning - Does it teach PM concepts while executing?
- Decision quality - At forks, does it present 3 context-aware options with a clear recommendation?
- Workload inversion quality - Does the assistant propose structure instead of pushing setup labor onto users?
- Persona fit - Are recommendations understandable and meaningful from the user/persona point of view?
Build prompts that adapt to user expertise level:
Based on your experience level:
- Beginner: [Guided questions with explanations]
- Intermediate: [Framework choices with rationale]
- Advanced: [Open-ended strategic queries]Create prompts that simulate different perspectives:
Now consider this from the perspective of:
- Engineering: [Technical feasibility questions]
- Design: [User experience considerations]
- Business: [Revenue and growth impact]
- Legal: [Compliance and risk factors]Build in continuous improvement:
After generating initial output:
1. "What assumptions should we validate?"
2. "What additional context would improve this?"
3. "What alternative approaches should we consider?"- Solves real PM pain you've experienced personally
- Teaches methodology through structural comments
- Adapts to context across industries and team structures
- Creates learning not just task completion
- Facilitates decisions well using one-question turns plus 3-option guidance
- Follow our architectural patterns - Study existing generators first
- Include rich pedagogical comments - Explain your design decisions
- Test across multiple scenarios - Validate with diverse PM challenges
- Focus on transferable skills - Help others learn to build, not just use
- Stakeholder alignment - Building consensus across functions
- Technical debt prioritization - Balancing features vs. infrastructure
- Pricing strategy - Market positioning and value-based pricing
- Team dynamics - Improving PM-Eng-Design collaboration
- Crisis response - Managing product emergencies and pivots
✅ Successfully used 5+ existing prompts
✅ Modified prompts for your specific context
✅ Created your first custom prompt using generators
✅ Tested across multiple AI platforms
✅ Built prompts integrating PM methodologies
✅ Created domain-specific prompt variations
✅ Designed prompts others successfully use
✅ Contributed improvements to existing generators
✅ Built meta-generators that create other prompts
✅ Helped others learn prompt building techniques
✅ Identified new architectural patterns
✅ Advanced the field through novel approaches
Every prompt you create teaches AI to be a better product management partner.
The product managers who thrive in the AI era won't be those who fear automation—they'll be those who teach machines to amplify human strategic thinking.
Your expertise + AI capabilities + structured methodology = Unstoppable product outcomes
Ready to become a prompt architect? Start with the universal builder, then specialize in your domain of expertise.