Smoke test
A smoke test is a small, fast set of tests that verify the most critical paths of a system work at all — does the app start, can a user log in, do the top three workflows respond. Smoke tests run on every deployment to catch broken builds before they reach users.
The name comes from electronics ('does it smoke when you plug it in?'). For software, smoke suites are typically 5-20 tests covering the happiest of happy paths. They're the cheapest insurance you can buy against a complete deploy disaster. Where smoke tests fail vs succeed: a passing smoke suite means the system isn't broken in obvious ways; it does NOT mean the system is bug-free.
Long-form posts that explore smoke test in depth — when to use it, common failure modes, how AI helps.
- Are AI-generated test cases worth shipping?Yes, with a sharp caveat — when they're tied to AC and reviewed by a human. Five categories where AI test generation is great, five anti-patterns to catch.9 min read
- Can AI write Gherkin? (yes — here's how)Yes. AI writes Gherkin well, often better than humans for surface area coverage. Five wins, five recognisable failure modes, and the prompts that work.8 min read
Related terms
- Regression test
A regression test verifies that previously working functionality still works after a code change.
- Integration test
An integration test verifies that multiple components work together correctly — a service hitting a real database, two microservices communicating, a frontend talking to a real API.