In Generative Copilot Studio Bots for Power CAT Copilot Studio Kit, the Topic Match Flows feature is not functioning.

Kavyasri Paidisetty 5 Reputation points
2025-04-01T08:04:10.7066667+00:00

Hi Everyone, I am working on the Power CAT Copilot Studio Kit. We configured the Copilot Studio Bot to both Classic Bot and Generative Bot, and imported the Copilot Studio Kit solution to PowerApps. We also created test suites and test cases. However, while the topic match flow worked in the Classic Bot, in the Generative Bot it is not working and shows "Not an exact match on the triggered topic or event name" even though the name is correct. Has anyone encountered this issue before? Any help would be greatly appreciated!

Microsoft 365 Developer Program
Microsoft 365 Developer Program
Microsoft 365: Formerly Office 365, is a line of subscription services offered by Microsoft which adds to and includes the Microsoft Office product line.Developer Program: A Microsoft program designed to help developers build cross-platform productivity experiences by giving them access to a developer subscription, sample data packs, documentation, training, experts, recommendations, and community events.
186 questions
0 comments No comments
{count} votes

2 answers

Sort by: Most helpful
  1. Petri Simolin 0 Reputation points Microsoft Employee
    2025-04-02T17:00:33.43+00:00

    Hello Kavyasri,

    With generative bots and generative responses, "Response Match" is not ideal test type as responses are non-deterministic, and with "Response match" the response needs to be exact the expected, or otherwise the test case will fail. "Generative answers" test type is more ideal for this purpose, as it uses AI Builder and LLM to validate the generative response.

    We are also releasing the next version of the Copilot Studio Kit very soon, and we have added support for additional comparison operators such as "contains" which might also work in your scenario. It allows checking if the response contains required keyword(s), ignoring everything else, and then passing or failing the test based on that. Great for templatized (dynamic) responses, and in some cases for generative responses as well.

    Hope this helps.


  2. Roshan Singh 0 Reputation points
    2025-04-09T04:25:23.8833333+00:00

    "The Topic Match Flows feature might not be working because it only functions properly in a production environment. In sandbox or development environments, conversation transcripts—which are required for topic matching—are not generated. To fix this, make sure Dataverse enrichment is properly configured and that you're testing in a production environment. Also, allow 30-60 minutes for the conversation transcripts to be created."


Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.