One system under test, every exercise
- User
- Cart
- Discount
- Fraud
- Payment
- Notification
The 120 minutes
Friday, 10 July 2026, Room R2, 12:15–14:15 CEST
- 12:150Intro, setup & tourActivate Copilot, clone the exercises repo and walk the checkout pipeline.15 min
- 12:30AUnit testsGenerate tests for calculateDiscount(). Weak prompts miss the three hidden bugs; strong prompts expose them.25 min
- 12:55BReview AI testsFind the tests that pass but shouldn't, then rewrite them with exact assertions.10 min
- 13:05CAPI testsSupertest coverage for the full checkout pipeline, with domain-rules.md as Copilot context.25 min
- 13:30DComponent & E2E testsReact Testing Library for StorePage, plus the Playwright checkout flow.20 min
- 13:50ECI guardrailsCoverage gates, a flaky test demo and copilot-instructions.md.10 min
- 14:00✓Wrap-up & Q&AThe trust framework, one action to take back, open discussion.15 min
Copilot has changed since the session. See what's new from July to October 2026 and how it affects each exercise.
Your progress
Saved in this browser as you mark steps complete and answer quiz checks.
Who it's for
- Software engineersSpeed up test creation with Copilot while keeping production-quality standards.
- QA & test automation engineersSee where GenAI tooling fits into existing automation workflows and quality gates.
- DevOps & platform engineersGovern AI-generated code and wire quality gates into CI/CD team-wide.
What you'll take away
- Where Copilot adds value in testing, and where human judgment is still essential
- Generate and refactor unit, API and integration tests responsibly
- Review AI-generated tests for false confidence, flakiness and security risks
- Define guardrails for quality, security and consistency across a team
- Bring AI-assisted testing into CI/CD without lowering confidence
- Roll out AI-assisted testing across engineering teams