We're setting up an AI agent for support requests, something internal facing for now, to test the capabilities. We would love to be able to point the agent to a consolidated history of all support chat history (a missive team inbox) in order for the agent to use that as knowledge base to learn form. Do you think that's something you would be able to expose? With LLMs being such a driver of innovation these days, I think that would be a very interesting move.