Test case generation
Produce in 15-30 minutes an exhaustive test plan (happy-path + edge cases) from a user story.
Test case generation is one of the most rentable activities to inject AI into the QA flow. From a user story, AI can produce in minutes 20-50 test cases covering expected behaviors, edge cases, errors. QA keeps central value: prioritize, execute, identify real bugs AI didn't think to test. This guide presents the workflow.
Step-by-step workflow
Submit user story and context
Story + acceptance criteria + technical context (API, UI, mobile). Richer context = more relevant generated cases.
Request 4 categories of cases
Happy-path (3-5 cases), edge cases (5-10), errors and invalid inputs (5-10), regression tests (3-5). Systematic coverage.
Hierarchize by priority
AI produces a lot; QA prioritizes. Criteria: business impact, usage frequency, criticality. Top 20% covers 80% of real bugs.
Convert to tool format
Per stack: Gherkin for Cucumber, TestRail/Xray format, or simply markdown list. AI can convert between formats.
Maintain over evolution
At each feature evolution: have test cases updated. Makes testing alive rather than debt.
Copyable prompts
2 tested and optimized prompts. Adapt the bracketed variables [VARIABLE] to your context.
Test plan from user story
You're a senior QA. Generate a test plan for this user story: **User story**: [STORY] **Acceptance criteria**: [LIST] **Technical stack**: [WEB / MOBILE / API] **Project context**: [USEFUL INFO] Produce: 1. **Happy-path** (3-5 cases): nominal behavior 2. **Edge cases** (5-10): boundary values, empty states, first uses, deactivated accounts, partial permissions 3. **Error cases** (5-10): invalid inputs, timeouts, network errors, concurrency conflicts, missing data 4. **Regression tests** (3-5): possible impact on existing features For each case: (a) ID, (b) title, (c) prerequisites, (d) steps, (e) expected result, (f) priority.
Targeted exploratory test
For this feature: [FEATURE] Produce a 60-90 min targeted exploration plan: 1. **Charters** (3-5): short exploration missions 2. **Tactics** per charter: approaches to use (boundary testing, error guessing, persona-based) 3. **Hidden risks**: zones where bugs are likely 4. **Questions to ask** during exploration Objective: structured but open exploration finding bugs automated tests don't.
Top tools for this use case
Curated selection of the 3 best AI tools for test case generation.
Why for this use case: Most rigorous for exhaustive case generation with well-anticipated edge cases.
Why for this use case: To generate in project context: code access, conventions, existing fixtures.
Why for this use case: Code Interpreter useful to generate varied test datasets.
Estimated ROI
Time saved
70% on planning (15-30 min vs 1-2h)
Quality gain
Exhaustive edge case coverage, format ready for QA tools
Stack cost
$20-30/month
Estimates based on 2026 benchmarks and user feedback. Actual ROI depends on your context.
Frequently asked questions
Are generated test cases sufficient?
For systematic coverage: yes. For creativity (improbable cases revealing subtle bugs): less. Best practice: AI for mechanical 80%, human exploration for remaining 20%.
Can AI prioritize test cases?
For indicative prioritization based on technical criticality: yes. For business prioritization (financial impact, affected client segment): less. QA arbitrates per context.
Should all generated cases be automated?
No. Classic rule: 70% automated (regression, smoke), 20% manual (exploration, UX), 10% out of scope. AI can advise distribution but team's choice.
Does AI really improve quality?
Indirectly: exhaustive coverage, fewer omissions. Also: frees time for exploration and critical testing. Net: fewer prod bugs, more release confidence.