Share
Report
85I have an agent in Copilot Studio. It has four knowledge sources—all Azure AI Search sources.
I often find that the agent chooses the escalation path because it seems to have an unknown indent.
But basically, it should send every message from the user to the four AI Search sources and build its result from that.
The only problem can arise if none of the four sources returns a result. But in my case, it fails even with simple terms or sentences where EACH of the four AI Search sources returns an answer.
How do I configure the agent so that it always queries the AI Search sources and returns results, even if they are poor? Do I do this via the instructions prompt? Or is there another way?
I have the same question (0)
All responses (
Answers (
