LangChain introduces ambient agents, a new approach to AI that reduces interaction overhead by responding to ambient signals and seeking user input only when necessary. The concept of ambient agents allows multiple agents to run simultaneously and incorporates human-in-the-loop interaction for notification, questioning, and review. LangGraph supports these agents with built-in persistence, human-in-the-loop support, long-term memory, and cron jobs. The system is demonstrated with an AI email assistant available as both a hosted service and an open-source project.
Table of contents
What is an ambient agent?Human-in-the-loopThe importance of human-in-the-loopAgent InboxWhy LangGraph is great for ambient agentsBuilding an AI email assistantSort: