The Evaluation trigger is used to test how your workflow handles a set of predefined inputs. Each test case runs the workflow independently, allowing you to compare AI responses, validate logic, or assess output quality across different scenarios.
Setup
In the Start node inspector, select Evaluation.
Test Cases
Define one or more test cases in the inspector. Each test case is a set of input values that gets fed into the workflow as if it were a real run. The workflow executes independently for each test case.
Use the results to:
- Compare how the workflow responds to different inputs
- Catch regressions when you change prompts or logic
- Evaluate AI response quality across a range of scenarios
When to Use
The Evaluation trigger is intended for development and quality assurance — not for production automation. Build and test your workflow with Evaluation, then switch to a production trigger (Manual, Schedule, Webhook, etc.) when you're ready to deploy.