Configuring agent memory

Configuring AI agent procedures manually is a massive hassle - it requires prompt engineering skills to optimize instructions for AI use, keeping track of the big picture across multiple memory files, and manually identifying and fixing conflicts. This is especially critical for sensitive policies like refunds, where you need the AI agent to follow instructions reliably.

You can use the Internal agent to help with memory updates by simply describing your procedures in chat or attaching documentation files.

The Internal agent to automatically search existing memory, create or update memory files, research historical tickets to find relevant test cases, and create test simulations - so you can validate changes before you publish them.