AI operations / Lab

Lab

Validate AI answers with simulations before rolling them out.

Creating a test case

1. Define the scenario, the evaluation criteria, and the first message.
2. When you run it, answers are generated through the same knowledge-search path as in production.
3. Continue the conversation in the right panel and check the answers.
4. Click Evaluate to get a pass/fail verdict with reasons, based on your criteria.

Each answer shows the knowledge it referenced and the folder instructions it applied, so you can trace why that answer came out.

Have more questions?

For SDK integration, see the Developer docs. For anything else, email hello@lillo.cx or use Contact in the console.