Hey everyone! My team has been working on a project that utilizes an AI agent along with AI search as its knowledge base. It was functioning pretty smoothly until recently. I've noticed that the agent just stopped pulling information from the knowledge source altogether. This issue seems to affect even older test agents that were working fine before; now they're hallucinating results instead of retrieving actual data. Has anyone else run into this problem? Is there a way to get the agent to actually conduct searches again? Also, I've noticed that even file-based sources seem to be failing. As an update, one of our developers had success after creating a new agent using the "New" AI foundry experience. Any thoughts or suggestions would be greatly appreciated!
3 Answers
How about asking CoPilot? I hear that it’s pretty good at tackling these kinds of issues and might have some insights.
Have you considered rebuilding the agent or reconnecting it to the knowledge base? Sometimes, resetting things can help clear up issues like this.
Just a heads up, creating a new agent seems to have worked for one of our team members! They used the "New" AI foundry experience and got better results.
Related Questions
Biggest Problem With Suno AI Audio
How to Build a Custom GPT Journalist That Posts Directly to WordPress