Planner Agent search tips

Understand query behavior, result limits, and how to get better matches.

Planner Agent search tips

Planner Agent uses meaning-based search to find relevant test cases.

It does not rely only on exact keyword matches.

How search works

When you describe a feature or journey, Planner Agent looks for related terminology and matching test coverage in your current project.

This helps it find useful results even when your wording differs from the wording used in existing test cases.

For example, a request for Login may also match related authentication flows.

A request for Fund Transfer may also match related transfer methods already described in the repository.

Choose a better query

Use a clear feature, journey, or release area.

Good examples include:

• UPI payments

• Payment checkout flow

• Password reset

Avoid vague prompts such as yes, ok, or short follow-up messages without enough context.

Result limits

Planner Agent lets you choose how many test cases to return before you run the search.

Use a lower number when you want a focused shortlist.

Use a higher number when you want broader plan coverage.

If the results feel too narrow, raise the limit and search again.

No-result searches

If nothing relevant is found, Planner Agent returns zero results instead of inventing test cases.

This usually means one of these things is true:

• the current project does not contain matching coverage

• the search should be rewritten with a clearer feature or journey name

• the search is being run in the wrong project

Common pitfalls

• Do not combine multiple unrelated requests into one search if you want separate result sets.

• Do not assume a missing result means the test case does not exist. It may be stored in a different project.

• Do not expect Planner Agent to create new test cases. It only finds and organizes existing ones.

Next step

Return to Planner Agent for the end-to-end workflow.