AI agent testing history and response comparison

Business problem

There is currently no way to track or compare AI agent responses across different testing sessions. When users update knowledge base sources or instructions, they cannot easily validate improvements because previous test results are lost, requiring manual documentation or screenshots to perform "before and after" comparisons.

Wished result

A visible history or log within the AI agent testing area that stores past questions and their corresponding answers. This would allow users to look back at previous iterations and verify if specific changes to the knowledge base have actually improved the AI's output.