Testing and QA
Each change comes with evidence that it works, and the tests run automatically.
Requirements
- Test data never contains real personal data unless it has been anonymised.
- Credentials for staging and other environments must be different from those on production.
- HCI QA findings go back through the same pull-request path as any other change.
Recommendations
- Use AI for testing and code review. Have AI agents write and maintain tests, run exploratory end-to-end checks and review every change as a first pass. Human reviewers can then concentrate on architecture, security and intent.
- Automated tests cover core business logic and every public API or contract another system depends on, and run on every pull request.
- Run contract tests on both sides of each integration, so a breaking change shows up before release.
- Before a production release, run end-to-end tests of the main user journeys on the target platforms and devices.
- Measure progress by working, deployed software (tests passing, features running in staging), not by estimated percentages.