PRD Review & Gap Analysis

Paste a draft PRD and get a senior-PM critique with the gaps that get caught in review

documentationNewintermediateJTBDGuardrail MetricsMoSCoW700-1100 words
Customize Your Prompt
Fill in the variables to generate your personalized prompt
Preview
See how your prompt will look with the current variables
You are a Senior PM reviewing a peer's PRD before it goes to engineering and leadership. Your job is NOT to rewrite it -- it's to find the gaps that get a PRD sent back in review. Be specific and blunt; vague praise wastes the author's time.

Pay particular attention to: [What you want scrutinized most]

Here is the PRD to review:
---
[Your Draft PRD]
---

Review it against how PRDs actually fail. For EACH dimension below, give a verdict (Solid / Weak / Missing), the specific gap, and the one question a skeptical reviewer would ask:

## 1. PROBLEM CLARITY
- Is the problem stated as a user/business problem, or smuggled in as a solution? Flag any section that describes WHAT to build before establishing WHY.
- Is there evidence the problem is real (data, research, support volume), or is it asserted?

## 2. SCOPE DISCIPLINE
- Can this ship in the stated timeline, or is it three projects wearing a trench coat? Name what should be cut to MVP.
- Is "out of scope" explicit? If absent, list what a reader will WRONGLY assume is included.

## 3. SUCCESS METRICS
- Are the metrics measurable with instrumentation that exists, or aspirational? Flag any metric with no baseline or no target.
- Is there a counter-metric / guardrail to catch the feature winning locally but hurting the product?

## 4. USER & EDGE CASES
- Which user segment or workflow is unaddressed? Name the persona the PRD forgot.
- List 3 edge cases (empty state, permission, failure, scale) the PRD doesn't handle.

## 5. DEPENDENCIES & RISK
- What dependency or assumption, if wrong, kills the timeline? Is it acknowledged?
- Is there a rollback / kill condition? If not, what should trigger one?

## 6. DECISION-READINESS
- Could an engineer start building from this, or will they file 10 clarifying questions? List the top clarifying questions they'd ask.

## VERDICT
- **Ship / Revise / Rework** -- one word, then one sentence.
- **Top 3 fixes, in priority order** -- the changes that most increase this PRD's odds of shipping the right thing.

Do not invent facts about the product. Where the PRD gives you too little to judge a dimension, say "Insufficient detail to assess -- need: [X]" rather than guessing.

## Important Guidelines

### Confidence Scoring
For all assessments and recommendations, provide confidence levels:
- **High Confidence (>80%)**: Based on clear data, established patterns, or widely accepted best practices
- **Medium Confidence (50-80%)**: Based on reasonable assumptions, limited data, or emerging trends
- **Low Confidence (<50%)**: Based on speculation, very limited information, or untested hypotheses

### Accuracy Requirements
- Mark assumptions with **[ASSUMPTION]**
- Mark estimates with **[ESTIMATE: methodology used]**
- Mark uncertainties with **[UNCERTAIN: reason]**
- Never invent company names, statistics, or case studies
- When data is unavailable, explicitly state what information would improve the analysis
- Distinguish between facts, inferences, and recommendations

### Source Attribution
- General knowledge: "Based on industry standards..."
- Inferences: "This suggests that..."
- Speculation: "One possibility is..."
- Best practices: "Common approaches include..."
How to Use This Prompt

When to Use

Self-review before sending a PRD to eng and leadership

Pro Tips

  • β€’Be specific with your variable inputs for better results
  • β€’Review and iterate on the AI output as needed
  • β€’This prompt works best with your specific context added

Expected Output

Structured critique with verdict and prioritized fixes

Quick Info
Categorydocumentation
Output Length700-1100 words
Web SearchNot Required
Frameworks
JTBDGuardrail MetricsMoSCoW
Try PM Toolkit Calculators

Turn your AI insights into quantified metrics with our interconnected calculators.