This is a concept build, not a client project. It shows how I would set up an AI voice agent for a real estate agency, using the same Vapi, Make and GoHighLevel pieces I've already shipped in my Vapi call-routing build.
The problem
Listing calls come in at the worst times: during showings, in the evening, at weekends. A buyer who reaches voicemail often calls the next listing instead. And when an agent does call back, they usually start from zero, without knowing which listing the caller wanted, their budget or their timeline.
Proposed workflow
Answer every listing call 24/7 with a voice agent that introduces the agency by name.
Capture name, phone number, which listing, and whether they're buying, selling or renting.
Ask the same qualifying questions to every caller: budget, timeline and pre-approval.
Book a showing or a valuation call on the agent's calendar, or hand the caller to the agent.
Send the lead and a call summary to GoHighLevel through a webhook, and trigger a follow-up if they didn't book.
Routing rules
Each type of call gets a different next step. Listing questions are answered only from the agency's listing information. Ready buyers get a showing slot, sellers get a valuation call, and anything the agent can't answer is passed to a person with the details already captured.
Fair housing and safety
The agent asks every caller the same questions and gives every caller the same listing facts. It doesn't comment on neighborhoods or who a home suits, and it never gives advice on offers, pricing or contracts. Those go to the agent.
Status
Concept only: this hasn't been deployed for a real estate agency yet, so there are no results to report. The webhook and CRM routing come from my live Vapi to GoHighLevel build, linked below. If you run an agency and want this built, see my real estate service.
Like this project
Posted Oct 26, 2026
Concept build, not client work: an AI voice agent for real estate that answers listing calls, qualifies buyers and sellers, and books showings.