Contextualizing chat responses based on conversation history
Topics: LLMO / GEO, Microsoft, Personalization, Probably in use, Search Intent, User Signals
This Microsoft patent describes a system that allows a chat agent, powered by an AI language model, to provide more personalized and relevant responses. It achieves this by analyzing a user’s conversation history in two main ways: by building a long-term user profile based on recurring topics and preferences, and by performing a semantic search of past conversations to find specific, granular details relevant to the user’s current query. The goal is to create a more contextual and satisfying chat experience by tailoring responses to the individual user.
