Build a copy-paste prompt library for my product-management work. Tailor it to my actual product and operating constraints, not generic PM advice. Product, users, and business model: [PRODUCT CONTEXT] My role, team, and decision rights: [ROLE AND TEAM] Current goals, metrics, and time horizon: [GOALS AND METRICS] Current roadmap, problem areas, and known evidence: [WORKING CONTEXT]
Tools, data access, and constraints: [TOOLS AND CONSTRAINTS]
Ask up to 3 clarifying questions only if a required input is missing. Then provide 12 distinct prompts, each 90-140 words, written for ChatGPT, Claude, or Gemini. Cover: problem framing, interview planning, transcript synthesis, opportunity assessment, PRD critique, experiment design, metric diagnosis, prioritization, roadmap trade-offs, stakeholder update, launch readiness, and retrospective.
For each prompt, include a title, when to use it, the exact prompt, and the inputs I must replace. Each prompt must demand a concrete artifact: for example, an opportunity-solution tree, interview guide, evidence table, assumptions log, PRD gap list, experiment brief, metric tree, RICE comparison with uncertainty notes, decision memo, or launch checklist.
Build in PM discipline: distinguish facts from hypotheses; identify the target user and job before proposing features; separate leading from lagging metrics; call out dependencies, edge cases, and risks; and avoid treating a prioritized list as a committed roadmap. Do not invent customer evidence, market data, engineering estimates, or stakeholder agreement. Before answering, check that every prompt has a defined audience, decision, and output format, and that no prompt asks the model to fabricate evidence.Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [PRODUCT CONTEXT] | Describe the product, target users, business model, and stage of the product. | B2B appointment scheduling software for independent dental practices; subscription product with 2,400 paying locations. |
| [ROLE AND TEAM] | State your role, partner functions, team size, and what decisions you own. | Senior PM for patient communications; one designer, eight engineers, data analyst shared across three teams; I own discovery and roadmap proposals. |
| [GOALS AND METRICS] | List current objectives, key metrics, baselines if known, and planning horizon. | Increase confirmed appointments by 8% this quarter; reduce no-shows; track confirmation rate, opt-out rate, and support tickets. |
| [WORKING CONTEXT] | Summarize active initiatives, user problems, research findings, and decisions in progress. | Practices want fewer manual reminder calls; interviews show patients dislike repeated SMS messages; deciding between two-way texting and smarter reminder timing. |
| [TOOLS AND CONSTRAINTS] | List available research, analytics, delivery tools, and constraints such as compliance or engineering capacity. | Amplitude, Zendesk, Dovetail, Jira, and Figma; HIPAA requirements; one major release per quarter. |
How to use
- Paste the completed prompt with a current product brief, rather than a generic company description.
- Save the resulting prompts that match decisions you make repeatedly and remove tasks outside your scope.
- For each AI output, check the evidence table and assumptions log before sharing it with your team.
- Follow up with: “Adapt the experiment-design prompt to our two-way texting decision and include guardrail metrics for opt-outs and support volume.”
Variations
PRD review
Use this before circulating a product requirements document.
Act as a critical product review partner. Review this PRD: [PRD]. The target user and job are [USER AND JOB], while constraints are [CONSTRAINTS]. Produce a gap table covering problem evidence, scope, non-goals, user flows, edge cases, metrics, instrumentation, dependencies, rollout, and support readiness. Label each issue blocking, important, or optional, and write a question that would resolve it. Do not add requirements unsupported by the supplied context. Flag where the document confuses a solution with the underlying problem.
Roadmap trade-off
Use this when several credible initiatives compete for one planning window.
Help me write a decision memo comparing these initiatives: [INITIATIVES]. Our objective is [OBJECTIVE], available capacity is [CAPACITY], and evidence is [EVIDENCE]. Create a comparison table with target segment, expected mechanism, confidence, effort range, dependencies, risks, opportunity cost, and measurement plan. Then recommend a sequence only as a conditional proposal, naming the assumptions that would change it. Do not invent revenue impact, engineering estimates, or customer demand. End with a short stakeholder-ready decision memo.
Research synthesis
Use this after interviews or usability sessions are complete.
Synthesize these research notes or transcripts: [RESEARCH NOTES]. The study question is [STUDY QUESTION], participant criteria are [PARTICIPANT CRITERIA], and product context is [PRODUCT CONTEXT]. Produce a findings report with evidence-backed themes, representative excerpts labeled by participant, counterexamples, unmet needs, opportunity areas, and unanswered questions. Separate direct observations from interpretation. Do not calculate prevalence from a small qualitative sample or invent quotes. End with testable product hypotheses and the next research activity for each.
Tips
- Use a prompt to produce a decision artifact, not a feature list; a metric tree or assumptions log makes weak reasoning visible.
- Give the model excerpts or exports from your actual research and analytics, then ask it to preserve participant IDs and distinguish quotes from interpretation.
- Require guardrail metrics in experiment prompts; a conversion lift can hide increased churn, opt-outs, support load, or inequitable outcomes.
- Keep uncertain effort estimates as ranges and label their source, since AI cannot replace engineering sizing or dependency discovery.
FAQ
Which product-management tasks are safe to use AI for?
Drafting research guides, structuring evidence, critiquing documents, and preparing decision options work well. Keep customer data and sensitive information within your company’s approved tools and policies.
Can AI prioritize my roadmap?
It can organize trade-offs and expose assumptions, but it cannot know your strategic commitments, technical dependencies, or customer promises unless you provide them.
How do I stop AI from producing generic PM advice?
Provide the target user, the decision to make, available evidence, constraints, and required artifact. Ask it to mark unsupported claims rather than filling gaps.