You are a product strategist.
Build an opportunity solution tree for the context below.
Context:
- Business outcome: [outcome]
- Target users: [users]
- Known opportunities or user problems: [opportunities]
- Current ideas or concepts: [ideas]
- Constraints: [constraints]
Instructions:
1. Clarify the desired outcome first.
2. Organize user opportunities underneath that outcome.
3. Suggest plausible solution directions without pretending they are all equally strong.
4. Identify where evidence is weak and should be validated.
5. Keep the structure decision-useful rather than overcomplicated.
Output format:
- Outcome statement
- Opportunity tree
- Solution directions
- Evidence gaps
- Recommended next validation steps
Missing information policy:
Proceed with assumptions unless the outcome or user context is unclear.
Quality bar:
- Outcome-first.
- Honest about uncertainty.
- Practical for prioritization.
02
Write a One-Page Product Strategy
Create a one-page product strategy that clarifies users, value, focus areas, and tradeoffs.
You are a product strategist.
Write a one-page product strategy for the context below.
Context:
- Product: [product]
- Target users: [users]
- Business goals: [goals]
- Current challenges: [challenges]
- Constraints: [constraints]
Instructions:
1. Clarify the product's strategic focus and what it is not trying to do right now.
2. Define the user value clearly.
3. Recommend a small number of strategic priorities.
4. Call out tradeoffs and risks.
5. Keep the strategy short enough to align a team.
Output format:
- Strategy summary
- Users and value proposition
- Strategic priorities
- Tradeoffs and risks
- What not to focus on
Missing information policy:
Proceed with assumptions unless the product or goals are unclear.
Quality bar:
- Concise.
- Opinionated.
- Useful for alignment.
03
Decide What to Include in an MVP
Decide what belongs in an MVP by focusing on the smallest credible path to user value.
You are a product scope strategist.
Decide what should be included in the MVP below.
Context:
- Product idea: [product idea]
- Target users: [users]
- Candidate features: [features]
- Constraints: [constraints]
- Launch goal: [goal]
Instructions:
1. Identify the core user problem the MVP must solve.
2. Sort features into must-have, support, later, and likely unnecessary.
3. Explain why each inclusion or exclusion decision makes sense.
4. Call out where too much scope will dilute learning.
5. Keep the recommendation realistic for the stated constraints.
Output format:
- Core MVP thesis
- Feature decision table
- Scope risks
- What to defer
Missing information policy:
Proceed with assumptions unless the product idea or user target is unclear.
Quality bar:
- Ruthless about scope.
- User-value first.
- Designed for learning, not completeness.
04
Turn a Product Quarter into a Review Memo
Create a product review memo that explains what happened, what mattered, and what should change next.
You’ll needgoals for the quarterwork shippedresultsmisses or surprisesnext-quarter context
Prompt
You are a product strategy reviewer.
Turn the quarter below into a product review memo.
Context:
- Goals for the quarter: [goals]
- Work shipped or launched: [work shipped]
- Results and signals: [results]
- Misses, surprises, or unresolved issues: [misses]
- Next-quarter context or constraints: [next quarter context]
Instructions:
1. Summarize what the team set out to do and what actually happened.
2. Distinguish output from meaningful outcome.
3. Identify what the team should learn from wins, misses, and mixed results.
4. Highlight tradeoffs, judgment calls, and open questions for the next quarter.
5. End with concrete recommendations rather than a retrospective summary only.
Output format:
- Executive summary
- What happened
- Outcomes versus outputs
- Key lessons
- Open questions and recommendations
Missing information policy:
Proceed with assumptions unless the quarter goals or results are too vague to evaluate.
Quality bar:
- Be honest about what was learned.
- Focus on decisions, not storytelling.
- Separate signal from status reporting.
05
Turn a Workflow Pain Point into a Lightweight AI Product Spec
Create a lightweight product spec for an AI workflow assistant based on a concrete job, user pain, and operating constraints.
You are a product strategist scoping an AI workflow tool.
Create a lightweight product spec using the context below.
Context:
- Workflow pain point or repetitive job: [pain point]
- Target users: [users]
- Current process: [current process]
- Desired outcome: [outcome]
- Constraints or risks: [constraints]
- Systems, tools, or data involved: [systems]
Instructions:
1. Define the job-to-be-done and the failure of the current workflow.
2. Scope the smallest useful AI-assisted product behavior that could help.
3. Clarify what the product should automate, suggest, or leave to human judgment.
4. Identify key risks around trust, data quality, or workflow adoption.
5. Keep the spec lean enough for a real prototype decision.
Output format:
- Product opportunity summary
- User and workflow context
- Proposed AI-assisted behavior
- Core requirements and boundaries
- Risks and open questions
Missing information policy:
Proceed with assumptions unless the workflow pain point or target user is unclear.
Quality bar:
- Concrete workflow grounding.
- Clear product boundaries.
- Prototype-ready scope.
06
Create a Lightweight PRD for an Internal AI Assistant
Create a lightweight PRD for an internal AI assistant with workflow scope, boundaries, risks, and success criteria.
You’ll needassistant use casetarget userscurrent workflowsuccess outcomeconstraintssystems or data involved
Prompt
You are creating a lightweight PRD for an internal AI assistant.
Use the context below.
Context:
- Assistant use case or job-to-be-done: [use case]
- Target users: [users]
- Current workflow or process: [workflow]
- Desired success outcome: [outcome]
- Constraints or risks: [constraints]
- Systems, tools, or data involved: [systems]
Instructions:
1. Define the assistant's narrow, high-value job clearly.
2. Clarify what the assistant should do, what it should suggest, and what should remain human-owned.
3. Identify boundaries around trust, permissions, and failure handling.
4. Keep the PRD lean enough for a realistic prototype decision.
5. Avoid turning the assistant into an overgeneralized product fantasy.
Output format:
- Product opportunity summary
- User and workflow context
- Core assistant behavior
- Requirements and boundaries
- Risks and success criteria
Missing information policy:
Proceed with assumptions unless the use case or target users are unclear.
Quality bar:
- Narrow, credible scope.
- Clear human-in-the-loop boundaries.
- Prototype-ready definition.
07
Evaluate Whether a Workflow Should Be Built, Bought, or Prompted
Evaluate whether a workflow should be solved by building software, buying a tool, or using prompts and operating process.
You are evaluating how a workflow problem should be solved.
Use the context below.
Context:
- Workflow problem or pain point: [problem]
- Current pain or cost: [pain]
- Frequency or scale of the problem: [frequency]
- Constraints: [constraints]
- Available tools or current stack: [tooling]
- Decision criteria: [criteria]
Instructions:
1. Assess whether the problem is best solved by building software, buying a tool, or using prompts plus operating process.
2. Compare the options on speed, flexibility, cost, control, and risk.
3. Identify where a prompt-based solution is sufficient versus fragile.
4. Flag hidden integration or adoption risks.
5. End with a clear recommendation and rationale.
Output format:
- Problem framing
- Build vs buy vs prompt comparison
- Best recommendation
- Risks and caveats
- What to validate next
Missing information policy:
Proceed with assumptions unless the workflow problem or decision criteria are unclear.
Quality bar:
- Clear comparison logic.
- Honest tradeoff handling.
- Recommendation tied to workflow reality.
08
Spec a Meeting Notes Agent with Human-in-the-Loop Boundaries
Spec a meeting notes agent with core behaviors, review boundaries, failure handling, and ownership rules.
You’ll needmeeting notes use casetarget userssource material qualityrequired outputstrust or compliance concernshandoff expectations
Prompt
You are specifying a meeting notes agent for a product team.
Use the context below.
Context:
- Meeting notes use case: [use case]
- Target users: [users]
- Source material quality or limitations: [source quality]
- Required outputs: [outputs]
- Trust, compliance, or privacy concerns: [concerns]
- Handoff or review expectations: [handoff]
Instructions:
1. Define the core job of the agent and its expected outputs.
2. Clarify what the agent should do automatically versus what should require human review.
3. Identify likely failure modes around misattribution, omission, and overconfidence.
4. Include workflow boundaries that make the agent safer to use in real teams.
5. Keep the spec concrete enough to build or test.
Output format:
- Agent job definition
- Core behaviors and outputs
- Human-in-the-loop boundaries
- Failure modes and safeguards
- Open implementation questions
Missing information policy:
Proceed with assumptions unless the use case or required outputs are unclear.
Quality bar:
- Concrete workflow boundaries.
- Failure-aware design.
- Testable behavior definition.
09
Review Whether an AI Assistant Scope Is Too Broad to Ship
Review whether an AI assistant scope is too broad to ship and identify a tighter v1 boundary.
You’ll needcurrent assistant concepttarget usersjobs to be doneproposed featuresriskslaunch constraints
Prompt
You are reviewing whether an AI assistant scope is too broad to ship.
Use the context below.
Context:
- Current assistant concept: [concept]
- Target users: [users]
- Jobs to be done: [jobs]
- Proposed features or behaviors: [features]
- Risks or concerns: [risks]
- Launch constraints: [constraints]
Instructions:
1. Assess whether the scope is coherent enough for a useful first launch.
2. Identify where the concept has expanded into multiple products or unclear promises.
3. Recommend the tightest v1 boundary that could still deliver real value.
4. Clarify what should be deferred or cut for now.
5. Keep the review grounded in product learnability and shipping reality.
Output format:
- Scope assessment
- Signs the scope is too broad
- Recommended v1 boundary
- Features to defer or cut
- Risks if the current scope remains
Missing information policy:
Proceed with assumptions unless the assistant concept or target users are unclear.
Quality bar:
- Shipping realism.
- Clear v1 boundaries.
- Honest scope reduction.
10
Define What an AI Assistant Should Refuse, Escalate, or Require Review For
Define what an AI assistant should refuse, escalate, or require review for, with workflow boundaries and risk logic.
You’ll needassistant roleworkflow contextrisk areasuser expectationscompliance or trust concernsfailure consequences
Prompt
You are defining what an AI assistant should refuse, escalate, or require review for.
Use the context below.
Context:
- Assistant role or core jobs: [assistant role]
- Workflow context: [workflow]
- Risk areas or likely failure types: [risks]
- User expectations: [expectations]
- Compliance, privacy, or trust concerns: [concerns]
- Failure consequences if it acts incorrectly: [consequences]
Instructions:
1. Define what the assistant can do safely, what it should route upward, and what it should refuse entirely.
2. Distinguish low-risk automation from actions that need a human decision.
3. Clarify the logic behind refusal, escalation, and review boundaries.
4. Keep the rules realistic for the product experience, not just theoretically safe.
5. Surface any obvious edge cases that need special handling.
Output format:
- Assistant boundary summary
- Allowed actions
- Review-required actions
- Escalate or refuse actions
- Edge cases and cautions
Missing information policy:
Proceed with assumptions unless the assistant role or main risk areas are unclear.
Quality bar:
- Explicit safety boundaries.
- Workflow realism.
- Clear escalation logic.
11
Define Success Metrics for a New Internal Workflow Tool
Define success metrics for a new internal workflow tool, including usage, trust, speed, quality, and business effect.
You’ll needtool purposetarget usersworkflow it changesdesired outcomesrisksmeasurement constraints
Prompt
You are defining success metrics for a new internal workflow tool.
Use the context below.
Context:
- Tool purpose or job-to-be-done: [tool purpose]
- Target users: [users]
- Workflow the tool changes: [workflow]
- Desired outcomes: [outcomes]
- Risks or known adoption concerns: [risks]
- Measurement constraints: [constraints]
Instructions:
1. Define success metrics that capture whether the tool is actually improving work.
2. Include adoption, speed, quality, trust, and workflow outcomes where relevant.
3. Distinguish leading indicators from lagging proof.
4. Avoid vanity measures that would look good without real workflow improvement.
5. Keep the framework practical for a team to monitor.
Output format:
- Success-metric framework
- Leading indicators
- Lagging indicators
- Anti-vanity cautions
- Review guidance
Missing information policy:
Proceed with assumptions unless the tool purpose or desired outcomes are unclear.
Quality bar:
- Workflow improvement focus.
- Balanced metric set.
- Low vanity-metric risk.
12
Define Handoff Rules Between a User and an AI Workflow Agent
Define handoff rules between a user and an AI workflow agent with triggers, responsibilities, and recovery paths.
You are defining handoff rules between a user and an AI workflow agent.
Use the context below.
Context:
- Workflow being shared between user and agent: [workflow]
- Agent role or tasks: [agent role]
- User role or responsibilities: [user role]
- Risk moments or likely failure points: [risks]
- Current handoff pain points: [pain points]
- Recovery or re-entry expectations: [recovery]
Instructions:
1. Define when the agent should take over and when it should return control.
2. Clarify what must remain explicitly owned by the user.
3. Include rules for low-confidence or ambiguous situations.
4. Design recovery paths when the handoff goes wrong.
5. Keep the rules concrete enough to implement and test.
Output format:
- Handoff model summary
- Agent-to-user handoff rules
- User-to-agent handoff rules
- Low-confidence and recovery rules
- Risks or edge cases
Missing information policy:
Proceed with assumptions unless the workflow or roles are unclear.
Quality bar:
- Explicit handoff boundaries.
- Recovery-aware design.
- Implementation usefulness.
13
Break a Workflow into Tasks Suitable for Agent Assistance
Break a workflow into tasks suitable for agent assistance, with human-owned work, risks, and sequencing logic.
You are breaking a workflow into tasks suitable for agent assistance.
Use the context below.
Context:
- Workflow to decompose: [workflow]
- Desired outcome: [outcome]
- Human roles involved: [roles]
- Risk areas: [risks]
- Task dependencies or ordering: [dependencies]
- Quality expectations: [quality]
Instructions:
1. Decompose the workflow into logical tasks and subtasks.
2. Identify which tasks are suitable for agent assistance and which should remain human-owned.
3. Clarify sequencing, dependencies, and error-sensitive points.
4. Include reasoning for why certain tasks are or are not good agent candidates.
5. Keep the decomposition practical for product design.
Output format:
- Workflow decomposition
- Agent-suitable tasks
- Human-owned tasks
- Risk-sensitive steps
- Sequencing and dependency notes
Missing information policy:
Proceed with assumptions unless the workflow or desired outcome are unclear.
Quality bar:
- Practical decomposition.
- Clear human/agent boundary.
- Risk-aware task mapping.
14
Compare Internal Tool Directions from a Workflow Perspective
Compare internal tool directions from a workflow perspective, including leverage, complexity, trust, and adoption risk.
You are comparing internal tool directions from a workflow perspective.
Use the context below.
Context:
- Tool options or directions being compared: [options]
- Workflow context: [workflow]
- User groups involved: [users]
- Adoption concerns: [adoption concerns]
- Constraints: [constraints]
- Decision criteria: [criteria]
Instructions:
1. Compare the options based on workflow leverage, usability, trust, and complexity.
2. Distinguish feature appeal from actual operational value.
3. Surface where one direction creates hidden adoption or integration cost.
4. Recommend the strongest direction and why.
5. Keep the comparison useful for a real product decision.
Output format:
- Comparison summary
- Workflow strengths and weaknesses by option
- Adoption and trust considerations
- Best recommendation
- Risks or open questions
Missing information policy:
Proceed with assumptions unless the options or workflow context are unclear.
Quality bar:
- Workflow-grounded comparison.
- Adoption realism.
- Clear recommendation logic.
15
Frame a Product Problem Clearly Before Solutioning
Frame a product problem clearly with user segment, evidence, cost of inaction, constraints, and decision-oriented next questions.
You’ll needsituation or signaluser segmentevidencebusiness impactconstraintsopen questions
Prompt
You are framing a product problem clearly before solutioning starts.
Use the context below.
Context:
- Situation, signal, or observed issue: [situation]
- User segment affected: [users]
- Evidence available: [evidence]
- Business impact or cost of inaction: [impact]
- Constraints shaping the space: [constraints]
- Open questions or uncertainty: [questions]
Instructions:
1. Distill the core user problem into a tight statement rather than a vague theme.
2. Clarify who is affected, how strongly, and how we know the problem is real.
3. Surface what happens if the team does nothing.
4. Separate known facts from assumptions or missing evidence.
5. End with the most decision-useful questions before solution design.
Output format:
- Problem-framing summary
- Core problem statement
- Evidence and impact
- Constraints and unknowns
- Next questions before solutioning
Missing information policy:
Proceed with assumptions unless the situation or user segment are unclear.
Quality bar:
- Tight problem definition.
- Evidence discipline.
- Decision usefulness.
16
Review a Draft PRD for Gaps Before Team Review
Review a draft PRD for missing context, weak success metrics, edge-case gaps, and cross-functional review risks.
You are reviewing a draft PRD before cross-functional team review.
Use the context below.
Context:
- Draft PRD: [prd]
- Product or feature context: [context]
- Review audience: [audience]
- Known uncertainties or unresolved areas: [uncertainties]
- Timeline pressure or scope constraints: [timeline]
- Main decision this PRD should support: [decision]
Instructions:
1. Review the PRD for missing context, vague requirements, weak metrics, and unaddressed edge cases.
2. Flag where the document may create confusion for design, engineering, or stakeholders.
3. Distinguish critical gaps from nice-to-have improvements.
4. Recommend the highest-value revisions before review.
5. Keep the critique practical and PM-quality rather than generic documentation advice.
Output format:
- PRD review summary
- Critical gaps
- Secondary improvements
- Review-risk notes
- Recommended next revisions
Missing information policy:
Proceed with assumptions unless the draft PRD or review audience are unclear.
Quality bar:
- Cross-functional review realism.
- Clear gap prioritization.
- Practical revision guidance.