Improve Cost and Quality through Testing

After the tests are stable, ask the code assistant to compare model choices for selected workflow nodes. It can test alternatives without changing the saved production workflow, then recommend a model placement only when the relevant end-to-end tests continue to pass.

In the Opportunity Manager sessions, this approach reduced reported AI Units while keeping all tests passing: the intent classifier stayed on a premium model, while several formatting and revenue-classification nodes moved to a basic model.