Rating
- size:
1 - complexity:
1
Purpose
Practice using AI to expand edge-case thinking without outsourcing the reasoning about what should actually be tested.
Scenario
A team owns a shipping-fee rule with these business rules:
- standard shipping is free at
$50or above - oversized items always add a
$15surcharge - premium members get free standard shipping below
$50, but not the oversized surcharge - digital-only orders should never have shipping fees
Task
Use AI to help create a small test matrix for the rule.
Then verify the matrix manually against the written rules.
Expected output
8-12candidate test cases- a short note explaining which cases came from AI
- a short note explaining which cases were added or corrected manually
Good AI uses
- asking for missed boundary conditions
- asking for contrasting examples
- asking for a test table format
Verification focus
- every proposed case must trace back to a rule
- every expected outcome must be manually checked
- duplicated or low-value cases should be removed
Anti-patterns to watch
- accepting AI-generated expected outcomes without checking the rules
- generating many cases and assuming quantity means coverage
- forgetting combinations such as premium plus oversized
Debrief prompts
- which good case did AI surface that you missed
- which wrong case did AI suggest confidently
- what did manual reasoning catch that the model did not