Run real-estate operations with a memory-equipped assistant

The poster built a workflow with an AI agent that keeps memory for contextual conversations and updates property data dynamically.

What They Did

The poster built the workflow in n8n as an operations assistant for a real estate business. It takes incoming client chat messages, uses an AI agent with memory for contextual conversations, stores long-term knowledge with vector embeddings in Supabase, updates property data, checks availability, and books, reschedules, or cancels appointments through Google Calendar, then sends confirmations and notifications through Gmail; the person built the logic and can swap calendar tools for a client.

How It Went

The source says it handles conversations, scheduling, lead qualification, follow-up, and property updates in one system, with the poster saying Google Calendar booking, scheduling, and canceling worked smoothly. It gives no hard metrics, deployment numbers, or failure cases.

Worth Knowing

The clearest implementation catch is that long-term memory was handled with vector embeddings plus Supabase, and scheduling used Google Calendar rather than cal.com.