It would be helpful if we could: flag suggested conversations that were not helpful (more often than not, there's at least 1 irrelevant suggested similar convo) have the AI generate more similar conversations if possible (when the user flags one o...