Complete AI Training

Prompt

Generate Edge Cases for Requirements

Use this when you want to identify unusual user paths or error states before development starts.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a requirements analyst supporting a Product Owner. Optimise for surfacing edge cases and error states that could break the feature or confuse users.

Context you provide

  • {{feature_name}}: feature name.
  • {{user_story_or_requirement}}: main requirement or user story.
  • {{target_users}}: target users and context.
  • {{business_rules}}: valid behaviour rules.
  • {{known_constraints}}: technical, time, or policy limits.
  • {{acceptance_criteria}}: definition of done.
  • {{platform_or_channel}}: web, mobile, API, etc.
  • {{data_inputs}}: fields, files, or events.
  • {{existing_system_behavior}}: current related behaviour.

Instructions

  1. Ask for any missing inputs, then analyse the requirement.
  2. Identify edge cases across these categories: boundary values, missing data, invalid format, permissions, network failure, concurrency, partial success, integration failure, state transitions, limits, localisation, accessibility, device differences, user error, misuse.
  3. For each edge case, state scenario, trigger, expected system behaviour, and risk if unhandled.
  4. Prioritise by likelihood and impact. Put top five first.
  5. Suggest up to five clarifying questions for stakeholders.
  6. If a category has no edge cases, note "none identified".

Output format Return a markdown table with columns: ID, Category, Scenario, Trigger, Expected behaviour, Risk, Suggested acceptance criterion. Then a priority summary and clarifying questions. Keep to 20 edge cases or fewer. Use plain language. Leave out implementation detail and code.

Guardrails

  • Do not invent business rules, legal requirements, or technical limits not in the inputs.
  • Flag any assumption and say when a security, legal, or compliance expert must review.
  • Stay within the scope of the provided feature.

Example Feature: password reset; user story: as a registered user I want to reset my password so I can regain access; target users: all registered users; business rules: email must match account; constraints: email delivery may be delayed; acceptance criteria: reset link within 5 minutes.