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.
Done withn8n
- 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.
Try it yourself with n8n
Build an n8n workflow with an AI agent that answers questions from [clients/tenants/team] about my [properties] using data in [spreadsheet or database]. Give it memory so it remembers each person's earlier conversation, and let it update property fields like [price, availability, status] when I tell it to. Before it changes any record or sends any message, show me what it will change and wait for my approval; stop once a test conversation works end to end.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: Reddit · May 1, 2026
Related use cases
Add todo items to Todoist from Telegram
A Telegram bot with an LLM and an orchestration platform (n8n or windmill) adds todos to the author's Todoist account from natural language messages.
Alert AI agencies to businesses with bad reviews as leads
n8n agents notify automation agencies when a business receives bad reviews so they can call first and offer a damage-control tool.
Manage customer enquiries from Telegram
An n8n system captures email and form enquiries, creates tasks, and lets the owner reply via Telegram while the AI sends emails and updates status.
Run a solo company with eight agent departments
A founder set up CEO, CFO, COO, lawyer, accountant, marketing, CTO, and improver agents that share memory and consult each other.
Schedule prospect meetings and log them in the CRM
A Lindy agent replied to prospect emails with time options, booked the chosen slot, logged it in HubSpot and sent a pre-call brief.
File roofing permits and inspection forms online
A roofing contractor had Grok Bot fill out and pull a permit, book an inspection through a captcha-protected form, and email two job-notification forms to his insurer.
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.