Improving the performance of your AI Agents

Hey! I’m Helena, part of the product team building out Compass - your tools in Front to measure and tune the quality of your AI Agents running in Front. 

We’ve released Opportunities - these will appear in your Agent Performance as fix suggestions (updates to knowledge, instructions, or tooling) to improve your AI agents based on its conversation history. 

To help us make them more useful, I'd love to understand how you work today:
 

1. How do you update your agent's knowledge, instructions, and tooling?

• Myself
• I delegate to teammates (which roles?)
• I collaborate with other teammates (which roles?)

2. Can you walk me through one fix you've made to your AI Agent?

• How did you know something needed to change?
• What was the change?
• How did you make it?

Thanks!