The most common question is about reaching the coverage required by the work teams… Using AI to create them just to improve coverage, will it always address the use cases we expect? How would it do that?
2 Likes
Using AI to improve coverage can be beneficial, but it does not guarantee that all expected use cases will be covered. AI can help identify patterns and generate test cases based on historical data, but it is crucial to complement it with human knowledge to ensure that critical scenarios and edge cases are covered. It is recommended to use AI as a support tool, not as a complete replacement for human analysis.
From my experience, AI has been very helpful for things like pure functions, validators, parsers, that kind of utilities where inputs and outputs are predictable. Also for basic CRUD and happy paths when I need quick coverage.
Where I do have to get my hands dirty manually is in error handling, edge cases, and business-specific things. There, AI won’t save you because it doesn’t understand the real context of your application.
Basically, I use it to get rid of the repetitive stuff and focus on the cases that really matter.
1 Like