Forum Discussion
Questions about Copilot Agents created using copilot studio, that use SharePoint & data-verse
I have developed many copilot agents using copilot studio, which have SharePoint sites and datavesrse tables as their source.. but i have those questions that i can not find a clear answers on them:-
1- Is there an order to how knowledge sources are indexed by a copilot agent? and if the source is SharePoint does Copilot use the classic SharePoint search ?
2- Why do we sometimes get inconsistent answers when asking the same questions?
3-Is there a way to integrate a PowerBI list instead of a SharePoint list?
4- Can the agent learn through user interaction? does generative AI need to be turned on for this to occur? In other words, if we have the same source, should we expect the same Copilot agent to provide more clear answers in the future, due to improvement to the underlying AI algorithms?
Thanks
2 Replies
A1: No, there is no specific order, and as far as I know, no new index is created. It simply uses the existing SharePoint index. If you enable Enhanced Search Results within the agent settings, the search becomes more accurate because certain features of Dataverse Search are activated. More details about Dataverse Search can be found here: https://learn.microsoft.com/en-us/power-apps/user/relevance-search-benefits and here: https://learn.microsoft.com/en-us/power-apps/user/relevance-faq
A2: In the test section of Copilot Studio, you can view the retrieved knowledge. Is this referencing a different file, or is it the same one? Do you have overlapping information across multiple documents that could be causing the system to produce such different answers?
A3: Not that I’m aware of.
A4: No, as of now, Copilot Studio uses a large language model (LLM) that is shared across multiple customers. This means nothing is stored or updated based on your specific interactions. Microsoft announced in 2024 that a form of "Memory" would be introduced, but I haven’t seen any updates beyond the initial announcement. You can read more about it here (search for “Memory and context”): https://www.microsoft.com/en-us/microsoft-copilot/blog/copilot-studio/microsoft-copilot-studio-building-copilots-with-agent-capabilities/ Also the screenshot shown in the Blog with the "History" section in Copilot Studio was never released.
Thanks for the reply.
A1) So for SharePoint , copilot uses the classic search index to find info?A2) No, on the same source such as SharePoint site, we can get different answers if we ask the same question twice, and most of the time the second answer will be more accurate.
A3) Clear thnks
A4) so even if we correct the agent, it will not learn from our feedback, is this what you mean?
