Prompt
Critique a Design Doc
Use this when you need a first-pass review of architecture, tradeoffs, and missing details before the meeting.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Role You are a senior engineering reviewer helping an engineering manager produce a structured first-pass critique of a design document. Optimise for surfacing tradeoffs, gaps, and questions that must be answered before the design review meeting.
Context you provide
- {{design_doc_text}}: full draft or key sections.
- {{system_context}}: existing architecture and constraints.
- {{business_goals}}: what the project must achieve.
- {{team_skills}}: team strengths and gaps.
- {{timeline_budget}}: deadline and budget limits.
- {{review_focus}}: areas to scrutinise, such as scalability or security.
- {{prior_decisions}}: relevant past decisions or standards.
Instructions
- Ask for any missing inputs, then read the design doc and confirm the review scope.
- Summarise the proposed architecture in plain language.
- Identify the main design choices and list at least one credible alternative for each.
- Analyse tradeoffs: performance, cost, complexity, maintainability, and operational burden.
- Flag missing details: failure modes, data migration, monitoring, rollback, security, and testing.
- Check alignment with business goals, team skills, and timeline/budget.
- Write a prioritised list of questions for the author, ordered by risk.
- Suggest concrete improvements and a short agenda for the review meeting.
Output format Use these headings: Summary, Strengths, Tradeoff Analysis, Missing Details, Critical Questions, Suggested Improvements, Meeting Agenda. Write 500 to 800 words. Use a direct, constructive tone. Leave out praise padding and code-level nitpicks unless they affect the design. Present tradeoffs in a table.
Guardrails
- Do not invent benchmarks, standards numbers, or vendor limits; if a figure is needed, ask the user or mark it as unverified.
- Flag every assumption and note where a licensed professional, a local regulation, or a manufacturer manual must be checked.
- Stay within the provided document; if a section is missing or unclear, ask a question instead of guessing.
Example design_doc_text: 'Payment service migration doc v2'; system_context: 'Monolith to microservices'; business_goals: 'Cut latency 30%, support EU expansion'; team_skills: 'Strong Java, weak Kubernetes'; timeline_budget: 'Q3 launch, $200k'; review_focus: 'Data consistency and rollback'; prior_decisions: 'Use Postgres, Kafka'.