Re: Traffic Management in Copilot Studio
So again, it will depend on volume whether we should have further interaction on this. Default throttles documented are around 800 requests per minute; if every one of your users used it twice a month for conversations of around 7 turns, you'd be in that range.
Do you have specifics around what channel you'd be using, how long the conversations might be, how frequently you expect those users to use the system? Also, usage of generative answers would be something to consider as well as that has significantly lower throttles due to scarcity of GPU's.