99+ ChatGPT Prompts for product managers

99+ ChatGPT Prompts Every Product Manager Needs in 2026

Unlock the full potential of AI-assisted product management with our comprehensive, production-grade library of prompts. From discovery and research to rollout and retrospectives, these prompts are meticulously crafted and tested against the latest AI models—designed to save you time, boost output quality, and sharpen decision-making across your product lifecycle.

Why Prompt Libraries Beat Ad-Hoc ChatGPT Use for Product Managers

Imagine a Series B SaaS product team where 14 product managers each spend roughly 90 minutes weekly reinventing similar ChatGPT prompts — drafting PRDs, synthesizing user interviews, performing competitive analysis, reflecting in sprint retrospectives, and updating stakeholders. None save or version their prompts, resulting in wildly inconsistent AI outputs and wasted effort.

This repeating inefficiency accumulates: 90 minutes per week translates into nearly 78 hours a year per PM, or about two full workweeks of overhead without producing direct business impact. This redundant effort slows team velocity and hampers output quality.

The solution? A rigorously maintained, versioned prompt library organized by product workflow stages and tested against leading AI models. This library empowers PMs to:

• Access high-quality, reusable prompts that cover core tasks end-to-end.
• Save significant time by eliminating repetitive prompt re-engineering.
• Standardize AI-generated artifacts for consistent stakeholder communication.
• Onboard new team members efficiently with documented, curated prompt sets.

Our prompts have been benchmarked on OpenAI GPT-5 series, Anthropic Claude models, and Google Gemini models. We recommend:

• GPT-5.4-mini (~$0.25/$2 per million tokens) for approximately 80% of routine product management activities.
• GPT-5.5 and Claude Opus 4.7 (higher cost) for executive communications, customer-facing content, and complex multi-constraint decisions.

Importantly, setting a workspace-level system prompt or custom instructions such as “You are advising a senior product manager at a 200-employee B2B SaaS company; provide specific, actionable recommendations with explicit trade-offs” cuts per-task prompt length by about 30% and sharpens AI output clarity substantially.

For deeper actionable insights on implementation patterns, see our related article: 99 Best ChatGPT Prompts for Product Managers to Master the Craft.

Discovery and Research Prompts (1–24)

Discovery and research are high-leverage areas where ChatGPT excels because it can efficiently detect patterns in unstructured user feedback, interviews, and market intel. Modern large language models (LLMs) with expansive context windows provide a powerful edge for synthesizing these insights.

For example, Claude Opus 4.7 outperforms GPT-5.4 on nuanced sentiment extraction and subtle user motivation cues in transcripts, while GPT-5.5’s massive 1.05-million token context window allows ingestion of 40+ raw interviews in one session — far beyond Claude’s 200,000 token limit. Choose the model that fits your data size and analysis complexity.

User Interview Synthesis Prompts (1–8)

1. Transcript clustering: “Below are 12 user interview transcripts from churned customers. Identify the five most frequently mentioned pain points. For each, quote two verbatim user statements, count occurrences across the corpus, and rate severity (1–5) based on emotional intensity. Output as a markdown table.”
2. JTBD extraction: “From the attached transcript, extract Jobs-To-Be-Done statements formatted as: ‘When [situation], I want to [motivation], so I can [expected outcome].’ List and rank by frequency.”
3. Anti-persona detection: “Analyze these eight interview transcripts. Identify which respondents are outside our target market, explaining why their needs diverge and flagging quotes to discount.”
4. Contradiction surfacing: “Find contradictions within transcripts where users’ stated preferences conflict with behaviors. Quote both sides.”
5. Question quality audit: “Review this interview script for leading or double-barreled questions and rewrite flagged ones neutrally.”
6. Follow-up generator: “Based on this transcript, list seven high-value follow-up questions to clarify ambiguities.”
7. Segment differentiation: “From 20 interviews across SMB, Mid-Market, and Enterprise segments, identify three problems unique to each and two universal pain points, output as a 3-column comparison.”
8. Recruitment screener: “Draft a six-question screener survey to recruit users matching this persona [paste persona], avoiding telegraphing correct answers.”

Survey Design and Analysis (9–16)

9. “Convert a list of research objectives into a survey of 12 or fewer questions mixing Likert, multiple-choice, and open-ended; estimate completion time.”
10. “Review a draft survey for bias, question order effects, and double-barreled items.”
11. “Analyze survey results (N=347) for statistically meaningful differences between power-users (top 10%) and median users.”
12. “Write a maximum two-question micro-survey validating a specific hypothesis, optimizing for response rate.”
13. “Draft NPS follow-up logic with verbatim prompts for promoters, passives, and detractors.”
14. “Thematize open-ended survey responses into 6–8 themes, with occurrence counts and sample quotes.”
15. “Identify and recommend combining redundant survey questions.”
16. “Write a stakeholder research brief summarizing goals, methodology, sample, and decision contexts, max 250 words.”

Competitive and Market Intelligence (17–24)

17. “Compare [Competitor A], [Competitor B], and our product across pricing, target segment, top 5 features, and public positioning, flagging inferences.”
18. “Analyze recent six release notes of a competitor to identify their strategic direction and prioritized customer problems.”
19. “Generate a SWOT analysis of [Competitor X] focusing on product capability gaps and adjacent market opportunities.”
20. “Draft a sales battlecard highlighting three strongest objection handlers and two honest weaknesses against [Competitor X].”
21. “Identify market white space from six competitor positioning statements; flag if absence is strategic or oversight.”
22. “Summarize top five user complaints and praises from 30 G2 reviews for [Competitor X].”
23. “Provide a 400-word summary of analyst coverage (Gartner, Forrester, IDC) on [market category], highlighting disagreements.”
24. “Infer future builds from recent [Competitor X] job postings.”

Pro Tip: Notice every prompt specifies an output format (e.g., markdown tables, ranked lists) paired with precise reasoning constraints (e.g., quote citations, inference flags). These details transform vague, generic responses into actionable research artifacts.

For practical examples and benchmarks, check out 15 Production-Ready ChatGPT System Prompts for Software Development Teams.

Strategy, Roadmap, and Prioritization Prompts (25–52)

Use these prompts to turn evidence, constraints, and business goals into explicit choices. Replace each bracketed field, and require the model to flag missing inputs.

Product Strategy Prompts (25–34)

25. “Given [product context], [customer evidence], and [business goals], identify the three most consequential strategic choices. Return a table with choice, evidence, trade-off, missing information, and recommendation.”
26. “Draft a concise product vision for [product] serving [target users]. Output a vision statement, the future user outcome, three guiding principles, and three things the product will not become.”
27. “Turn [research findings], [market context], and [company objective] into a strategy narrative. Structure the response as diagnosis, guiding policy, coordinated actions, and unresolved questions.”
28. “Translate [objective] into measurable user and business outcomes using [available baseline data]. Present each outcome with a definition, leading indicator, guardrail, and validation method.”
29. “Map [desired outcome] to [research evidence] and [candidate solutions]. Return an opportunity tree with evidence, key assumptions, and the riskiest branch.”
30. “Compare [segment A], [segment B], and [segment C] using [research], [commercial constraints], and [strategic criteria]. Produce a decision matrix, recommend one segment, and state what evidence could reverse the choice.”
31. “Review [positioning statement] against [target audience], [alternatives], and [evidence]. List unsupported claims, then provide a rewrite and rationale.”
32. “Develop three meaningfully different strategies for [goal] under [constraints]. For each, provide the core choice, expected benefit, central trade-off, earliest test, and stop condition.”
33. “Extract the assumptions embedded in [proposal]. Return a ranked register with assumption, category, supporting evidence, uncertainty, consequence if wrong, and suggested test.”
34. “Assume [strategy] failed after [time horizon]. Write a pre-mortem table containing five plausible causes, early warning signals, preventive actions, and owner placeholders.”

Roadmap Planning Prompts (35–43)

35. “Build an outcome-based roadmap from [objectives], [validated problems], [candidate initiatives], and [constraints]. Organize the output by outcome, evidence, initiative, dependency, and open assumption.”
36. “Place [initiatives] into now, next, or later using [criteria] and [capacity]. Return placement, rationale, dependency, and reconsideration condition.”
37. “Cluster [initiative list] into coherent roadmap themes tied to [product goals]. Output each theme with its intended outcome, included initiatives, success signal, and any unrelated work.”
38. “Inspect [roadmap] alongside [team responsibilities] and [known external commitments]. Produce a dependency map listing dependency, owner placeholder, affected work, sequencing effect, and mitigation.”
39. “Using [available capacity], [fixed commitments], and [candidate work], create conservative, balanced, and aggressive roadmap scenarios. Show included work, postponed work, assumptions, and principal risk for each.”
40. “Act as a skeptical roadmap reviewer. Based on [roadmap], [evidence], and [constraints], return ten challenge questions grouped under evidence, sequencing, capacity, and opportunity cost.”
41. “Classify each [roadmap item] as a commitment, forecast, experiment, or option. Rewrite labels that overstate certainty.”
42. “Prepare a roadmap-change memo for moving [initiative] from [old position] to [new position]. Use the supplied [new evidence] and [affected commitments] to produce a summary, rationale, impact, and next review point.”
43. “Compare [roadmap period] with [planned outcomes], [delivery record], and [learning]. Return a quarterly review organized as outcomes achieved, misses, changed assumptions, decisions, and proposed adjustments.”

Prioritization and Decision Prompts (44–52)

44. “Score [initiatives] against [criteria and weights] using only [supplied evidence]. Output the scoring table, confidence for each score, final ranking, and sensitivity notes.”
45. “Rank [opportunities] by impact, evidence strength, and effort using [research and estimates]. Mark absent inputs as unknown, then return a ranked table and a short explanation of the top choice.”
46. “For every item in [priority list], identify what selecting it would displace under [capacity limit]. Present a table with selected work, displaced work, affected outcome, and hidden opportunity cost.”
47. “Compare building, buying, and partnering for [capability] within [constraints]. Return a matrix covering control, implementation burden, ongoing ownership, switching difficulty, risks, and recommended next step.”
48. “Evaluate [options] against [decision criteria]. Separate verified facts from judgments in a decision matrix, then list the evidence that could reverse the recommendation.”
49. “Synthesize [stakeholder positions] about [decision]. Output the shared goal, points of agreement, conflicting assumptions, unresolved evidence, and a proposed decision process.”
50. “Assess whether [decision] is easy or difficult to reverse using [constraints and dependencies]. Return the smallest safe commitment, validation step, rollback trigger, and point of no return.”
51. “Triage [request backlog] using [strategy], [evidence], and [capacity]. Produce four lists—investigate, schedule, decline, and revisit—with a one-sentence reason and next action for every request.”
52. “Explain why [initiative A] ranks above [initiative B] based on [evidence], [criteria], and [constraints]. Format the response as decision, supporting reasons, acknowledged cost, uncertainty, and review trigger.”

Pro Tip: For roadmap and prioritization prompts, supply the evidence and scoring rules separately. Ask the model to expose unknowns, displaced work, and conditions that could change the recommendation.

Execution: PRDs, Specs, Stakeholder Comms (53–80)

For execution work, specify the audience, required decision, and acceptance conditions. Treat generated documents as drafts for cross-functional review.

Product Requirements Document Prompts (53–62)

53. “Create a product requirements document outline for [initiative] from [problem evidence], [target users], and [constraints]. Include problem, outcome, scope, requirements, exclusions, risks, dependencies, and open questions.”
54. “Rewrite [draft problem statement] using [user evidence] and [desired outcome] without prescribing a solution. Include affected users and remaining uncertainty.”
55. “Convert [research notes] and [workflow description] into requirements for [feature]. Return a table with requirement, rationale, priority category, evidence source, and acceptance condition.”
56. “Draft nonfunctional requirements for [feature] using [operating context] and [known constraints]. Organize them under reliability, accessibility, security, observability, and maintainability, marking every item that needs specialist confirmation.”
57. “Turn [workflow] and [user roles] into user stories. Include the user need, intended outcome, acceptance conditions, and unresolved assumption.”
58. “Write testable acceptance criteria for [requirement] from [business rules and examples]. Use a scenario table covering expected behavior, boundary conditions, one failure case, and recovery.”
59. “Audit [PRD] for scope creep against [intended outcome] and [release constraint]. Divide the output into essential work, optional enhancements, unrelated requests, and recommended exclusions.”
60. “Inspect [PRD] and [supporting specifications] for conflicting requirements, undefined terms, and unverifiable claims. Return an issue log with location, problem, consequence, and proposed clarification.”
61. “Generate an edge-case inventory for [workflow] using [states], [business rules], and [known failure modes]. Group the result by invalid input, interruption, permission, concurrency, and recovery.”
62. “Prepare a review agenda for [PRD] and [attendee roles]. Output the required decisions, disputed assumptions, pre-reading, discussion order, time-box placeholders, and decision-record template.”

Specification and Delivery Prompts (63–70)

63. “Specify [feature workflow] from entry to completion using [requirements and business rules]. Return numbered steps plus alternate paths, error states, recovery behavior, and unresolved questions.”
64. “Derive a state model for [entity] from [rules and lifecycle notes]. Output a transition table with current state, trigger, condition, next state, side effect, and prohibited transitions.”
65. “Extract data requirements from [feature description] and [reporting needs]. List each concept, purpose, source, validation need, and missing definition.”
66. “Compare [interface contract] with [product requirements] and [consumer expectations]. Return a compatibility review listing ambiguous fields, mismatches, failure behavior, migration risk, and decision needed.”
67. “Design an instrumentation plan for [feature] around [product questions] and [metric definitions]. Output essential events, trigger, properties, validation method, owner placeholder, and the decision each event supports.”
68. “Split [initiative] into independently testable release slices using [dependencies] and [learning goals]. Present each slice with user value, included scope, excluded scope, validation method, and sequencing rationale.”
69. “Build a delivery risk register for [initiative] from [plan], [dependencies], and [team concerns]. Include risk, trigger, potential impact, mitigation, contingency, and owner placeholder.”
70. “Create a readiness checklist for [release] using [scope], [support plan], and [measurement plan]. Group checks under product, engineering, operations, support, analytics, communications, and rollback.”

Stakeholder Communication Prompts (71–80)

71. “Write an executive update for [initiative] using [progress data], [risks], and [decision needed]. Format it as outcome, evidence, current status, material risk, recommendation, and requested action.”
72. “Transform [team notes] into a weekly status update for [audience]. Use sections for progress, outcome signals, blockers, decisions, and owner-tagged next actions.”
73. “Draft a decision memo for [decision] from [options], [evidence], and [constraints]. Include context, criteria, options, recommendation, strongest dissenting view, risks, and decision deadline.”
74. “Prepare a pre-read for [meeting] using [background materials] and [required decision]. Output an executive summary, evidence, options, unresolved questions, and the specific choice attendees must make.”
75. “Convert [meeting notes] into a structured record. Return decisions with rationale, action items with owner and due-date placeholders, deferred questions, risks raised, and follow-up meetings.”
76. “Draft a calm escalation about [issue] for [stakeholder]. Organize it as observed facts, possible impact, actions already taken, help needed, decision deadline, and next update point.”
77. “Communicate a delay to [release] using [changed facts], [mitigation], and [revised plan]. Produce a stakeholder-ready message with impact, rationale, current action, dependencies, and next update.”
78. “Decline [request] for [requesting audience] based on [evidence and constraints]. Write a respectful response that explains the decision, acknowledges the need, offers [alternative], and avoids unsupported commitments.”
79. “Create a cross-functional handoff brief for [receiving team] from [scope], [dependencies], and [acceptance conditions]. Include context, deliverables, owners, interfaces, risks, open questions, and completion criteria.”
80. “Rewrite [product update] for [audience] and [channel]. Preserve supplied facts and list any omitted details.”

Pro Tip: In requirements and communication prompts, require owner placeholders, acceptance conditions, and unresolved questions. This makes missing decisions visible before polished language creates false confidence.

Analysis, Launch, and Post-Launch Prompts (81–99+)

Define decision rules and guardrails before interpreting results. This reduces the temptation to reshape the analysis around a preferred outcome.

Product Analysis Prompts (81–87)

81. “Build a metric tree for [desired outcome] using [business model] and [user journey]. Return the primary outcome, contributing inputs, guardrails, known relationships, and assumptions requiring validation.”
82. “Define [metric] from [business rules], [data sources], and [decision context]. Output its purpose, formula, unit, eligibility rules, exclusions, reporting cadence, owner placeholder, and known limitations.”
83. “Analyze [funnel data] across [segments and period]. Present conversion by stage, the most consequential drop-off, data-quality warnings, plausible explanations, and the next diagnostic check.”
84. “Interpret [cohort table] in light of [product changes] and [measurement notes]. Separate observations, hypotheses, causal claims that cannot yet be supported, data concerns, and recommended follow-ups.”
85. “Design an experiment for [hypothesis] using [target population], [constraints], and [available metrics]. Return treatment, comparison, primary outcome, guardrails, assignment approach, stopping rule, and analysis plan.”
86. “Evaluate [experiment results] against [predefined decision rule] and [guardrails]. Output the observed result, uncertainty, data limitations, segment differences, recommendation, and conditions for another test.”
87. “Create an investigation plan for [metric anomaly] using [dashboard evidence], [recent changes], and [known incidents]. Prioritize checks for instrumentation, segment mix, seasonality, releases, and external factors in a table.”

Launch Planning Prompts (88–94)

88. “Create a launch plan for [feature] using [scope], [target users], [dependencies], and [success measures]. Return phases, deliverables, owner placeholders, entry criteria, risks, and communication points.”
89. “Design a staged rollout for [feature] across [eligible audience] under [operational constraints]. Output each stage with exposure criteria, evidence to collect, pause condition, rollback trigger, and approving owner placeholder.”
90. “Build a go/no-go review template for [release] from [mandatory requirements], [known risks], and [launch date]. Separate blocking criteria, accepted risks, evidence links, decision owners, and final disposition.”
91. “Draft launch messaging for [audience] using [value proposition], [confirmed capabilities], and [known limitations]. Provide a headline, benefit summary, limitation statement, and required action.”
92. “Prepare an internal enablement brief for [team] about [release]. Structure it as intended users, supported use cases, limitations, demo flow, common objections, escalation path, and reference materials.”
93. “Develop a support-readiness package for [release] from [known workflows], [failure modes], and [support channels]. Output likely questions, diagnostic steps, approved responses, escalation rules, and feedback tags.”
94. “Write a launch-day operating plan for [release] using [rollout schedule], [monitoring], and [team roles]. Include checkpoints, decision authority, communication channel, pause criteria, rollback procedure, and incident handoff.”

Post-Launch and Retrospective Prompts (95–99)

95. “Summarize [early usage data], [support reports], and [user feedback] for [release]. Return confirmed findings, weak signals, anomalies, data gaps, and the next observation window.”
96. “Prepare a post-launch report comparing [release results] with [intended outcomes and guardrails]. Organize it by adoption, outcome movement, incidents, user feedback, lessons, and recommended decisions.”
97. “Cluster [post-launch feedback] from [sources] into recurring themes. Output each theme with frequency indicator, representative evidence, affected segment, classification as defect or unmet need, and suggested owner.”
98. “Create a blameless retrospective for [initiative] using [timeline], [results], and [team observations]. Provide prompts for intended plan, actual events, surprises, missed signals, contributing conditions, and owner-tagged improvements.”
99. “Recommend whether to expand, revise, hold, or retire [feature] based on [outcomes], [costs], [feedback], and [constraints]. Return the decision, supporting evidence, strongest counterargument, next action, and review trigger.”

Pro Tip: Before supplying launch results, include the original success criteria, guardrails, and rollback rules. Ask the model to distinguish observations from explanations and to preserve contrary evidence.

Conclusion

The strongest prompts in this library pair a defined task with concrete inputs, an explicit output format, and constraints that prevent unsupported assumptions. Replace every bracketed field, preserve uncertainty, and review consequential outputs with the responsible specialists.

Keep the library versioned as workflows change. The “99+” label reflects this extensible format: add prompts beyond 99 as new workflows emerge while retaining the same structure of task, supplied evidence, reasoning boundaries, and a reviewable output format.