PM Prompt Library

25 battle-tested AI prompts for product managers. Copy, paste, customize, ship.

Works in ChatGPT, Claude, and any LLM. Each prompt tested across 50+ real PM scenarios.

These 25 cover the core PM workflow. The full library has 82 prompts across 12 categories, with advanced variants, chaining techniques, and domain-specific versions for B2B, B2C, marketplace, and AI products.

Get the full PM Prompt Library (82 prompts) →

---

What's Inside

Requirements

#	Prompt	Use When
1	User Story Generator	Turn vague requests into structured user stories with acceptance criteria
8	PRD Reviewer	Critique a PRD before sharing it with the team
25	API Spec Reviewer	Review an API spec from a product (not code) perspective

Strategy

#	Prompt	Use When
2	Competitive Teardown	Analyze a competitor's product, feature, or positioning
10	Roadmap Prioritizer	Prioritize a list of features using a structured framework
12	OKR Writer	Turn goals into properly structured OKRs
16	Pricing Page Reviewer	Get feedback on a pricing page before launch
21	Go-to-Market Planner	Plan a feature or product launch
23	Competitive Moat Analyzer	Analyze what makes a competitive advantage defensible

Execution

#	Prompt	Use When
3	Launch Risk Assessment	Stress-test a launch plan before shipping
9	Sprint Planner	Plan a sprint from a backlog of tickets and priorities
14	Postmortem Writer	Write a structured postmortem after something went wrong
15	Onboarding Flow Designer	Design or improve a user onboarding flow
17	Feature Flag Planner	Plan a feature flag rollout strategy
20	Bug Report Writer	Turn a vague user complaint into a clear engineering ticket

Analytics

#	Prompt	Use When
4	Metric Definer	Define success metrics for a feature or initiative
7	A/B Test Designer	Design an A/B test with hypothesis, variants, and sample size
18	Data Request Writer	Write a clear data request for your analytics team
24	Retention Analyzer	Diagnose why users are churning or not coming back

Communication

#	Prompt	Use When
5	Stakeholder Update	Write a status update fast
13	Release Notes Writer	Turn a changelog into user-facing release notes
19	Meeting Agenda Creator	Create a focused meeting agenda that respects everyone's time

Research

#	Prompt	Use When
6	User Interview Guide	Prepare for user interviews with a structured script
11	Customer Feedback Analyzer	Identify patterns in a pile of customer feedback
22	User Persona Builder	Build user personas from research data, not assumptions

How to Use

1. Open the prompt file
2. Copy the prompt
3. Replace the [BRACKETED] sections with your specifics
4. Paste into ChatGPT, Claude, or your LLM of choice

Every prompt includes:

• Full prompt text (ready to copy)
• When to use it
• Example input/output
• Tips for better results

Want More?

These 25 prompts are fully functional -- use them as-is. The full library of 82 prompts covers:

• PRDs and specs
• User research synthesis
• Competitive analysis
• Launch planning
• Stakeholder communication
• Metrics and analytics
• Roadmap planning
• Sprint planning
• Customer interviews
• A/B test design
• Post-mortems
• OKR writing

Get the full PM Prompt Library (82 prompts) →

Contributing

Found a bug, have a better variant, or want to add a new prompt? See CONTRIBUTING.md and start from prompts/TEMPLATE.md.

---

Built by Aakash Gupta | Product Growth Newsletter

---

Tips for Writing Better Prompts

Apply these whether you use prompts from this library or write your own.

1. Set the role and audience

Tell the LLM who it is and who the output is for. "You are a senior PM writing for an engineering team" produces vastly different output than no role at all.

2. Be specific about format

Not "give me a summary." Instead: "3-5 bullet points, each under 20 words, suitable for a Slack update." Tighter format constraints produce more usable output.

3. Provide context, not just instructions

Bad: "Write user stories for a calendar feature." Good: "Write user stories for a calendar feature in a B2B project management tool. Users are mid-market teams (50-200 people). The calendar should sync with Google Calendar and show project deadlines."

4. Use the "do / don't" pattern

Add explicit constraints after your main instruction:

• DO: use concrete metrics, cite the data I provided, flag assumptions
• DON'T: use jargon without defining it, make up statistics, exceed 2 pages

5. Ask for reasoning before conclusions

For analytical prompts (prioritization, trade-off analysis), require the LLM to show reasoning before the recommendation. This surfaces flawed logic early.

6. Iterate in the same conversation

First drafts are rarely final. Follow up:

• "Make the success metrics more specific"
• "Rewrite the problem statement with more urgency"
• "Add a section on risks"

Faster than crafting the perfect prompt upfront.

7. Include examples of good output

Show a snippet of what great looks like. LLMs pattern-match well -- even one example of the tone, depth, and format you want dramatically improves output.

8. Use placeholders consistently

Stick to [BRACKET CAPS] for user inputs: [PRODUCT NAME], [TARGET METRIC], [TIME PERIOD]. Keeps prompts scannable and reusable.