Copilot Agent Inconsistency in Teams

Question

A Microsoft Copilot Studio agent has been created and published into Teams. While the agent functions well within the Teams space, the generative answers it produces lack comprehensiveness and do not provide correct links to their sources. Has anyone encountered a similar issue with a disparity between the agent's performance in the Copilot Studio testing panel versus within the Teams chat? Efforts have been made to republish with more in-depth prompts, upload as a zip file instead of directly from the studio, and ensure the agent has the necessary access to the relevant channels but the agent is still not working properly. Any help would be appreciated

Answer 1

Hello Reuben Francis,

Thanks for reaching out. Could you please check below things to ensure the agent is working properly.
Make sure the prompts employed to train the agent are specific and well-defined. This can enhance the thoroughness of the responses. Update and validate data sources which the agent uses periodically. This will make sure the accuracy and validity of information supplied. Confirm that the agent is properly authorized and has access to all applicable channels and information in Teams. This can assist in ensuring that the agent is able to access and deliver the right information. Keep a regular check on the performance of the agent in both the testing panel and Teams.
Ref: https://learn.microsoft.com/en-us/microsoft-365-copilot/extensibility/declarative-agent-instructions
https://learn.microsoft.com/en-us/microsoft-365-copilot/extensibility/declarative-agent-capabilities-ids?tabs=explorer
https://learn.microsoft.com/en-us/microsoft-365-copilot/extensibility/copilot-studio-agent-builder-publish
Thanks, 

Smita

************************************************************************* 

If the response is helpful, please click on "upvote" button. You can share your feedback via Microsoft Copilot Developer Feedback link. Click here to escalate.