AI Real Estate AssistantBuilt on n8n + GPT Start a build

Lead capture · Qualification · Booking

Your next buyer messages at 11:40pm. Someone should answer.

We build AI assistants that reply the moment an enquiry lands, qualify the lead in conversation, match it against your live listings, and put the viewing straight into your calendar — on WhatsApp, your website, Messenger or SMS.

Maple & Co. Estates WhatsApp Business
21+ real estate teams live WhatsApp · Web · Messenger · SMS Books into your real calendar You keep the source code

Recorded walkthrough

This is the system running, not a mockup.

A full pass through the assistant: an enquiry arrives, the agent reads it, pulls matching properties, qualifies the buyer, and writes the booking back to the calendar and CRM — with the n8n workflow visible as it fires.

ai-real-estate-assistant — live run
  • The questions It asks only for what it still doesn't know — no form, no repetition.
  • The data Matches come out of the live listings sheet, not a hardcoded demo list.
  • The booking The slot is written back to the calendar and confirmed in the same thread.

21+

Real estate teams running an assistant we built — brokerages, independent agents and property managers.

24/7

Every enquiry answered on arrival, including the ones that land at 2am on a Sunday.

100%

Source code, workflows and prompts handed over. Nothing stays locked on our side.

  • Runs on your own accounts and API keys
  • Scope and price agreed before a line is built
  • Connects to the CRM you already use
  • Support continues after delivery

What it handles

The work that eats your evenings.

Answers every enquiry

Website chat, WhatsApp, Messenger, SMS — one assistant, one memory. It picks up the thread where the buyer left it, whatever channel they came back on.

Qualifies in conversation

Budget, timeline, financing, buy-to-live or buy-to-let. It asks like a person, scores the lead, and flags the ones worth calling first.

Matches your listings

Connected to your sheet, CRM or property API — so it recommends what you actually have today, at the price the buyer can actually clear.

Books the viewing

Checks the agent's real availability, holds the slot, sends the confirmation, and chases the no-shows without being asked.

Knows when to step back

Legal questions, offers, anything off-script — it hands over to a named human with the full transcript attached, instead of guessing.

Leaves a record

Every conversation lands in your CRM as a contact with notes, score and next action. Nothing lives only in the bot.

Under the hood

Six steps between "hello" and a booked viewing.

Built on n8n with GPT and Gemini doing the reasoning. Every step is a node you can open, read and change — no black box, no vendor lock-in.

  1. 01

    Enquiry arrives

    WhatsApp, web widget, Messenger or SMS webhook fires into the workflow.

  2. 02

    Intent read

    The model extracts what they want: area, beds, budget, urgency, and whether this is a new lead or a returning one.

  3. 03

    Listings matched

    Query against your Sheet, CRM or MLS feed. Only live inventory, only inside their range.

  4. 04

    Lead qualified

    Missing details asked for in natural conversation, then scored against the rules you set.

  5. 05

    Viewing booked

    Real availability checked, slot held, confirmation sent to buyer and agent.

  6. 06

    Written back

    Contact, transcript, score and next action pushed to your CRM. Agent gets the alert.

n8nOpenAI GPTGoogle GeminiPython · FastAPI WhatsApp Cloud APIGoogle CalendarSheetsHubSpot · GHL

Other systems shipped

Same pattern, different front desk.

AI hotel front desk

Takes the booking, verifies the payment screenshot against the invoice, and releases the room only when the numbers agree.

Booking · Payment verification

AI voice receptionist

Picks up the call, understands the request, reads the calendar out loud, and books the appointment before hanging up.

Voice · Scheduling

Property search agent

Search and qualification platform for a brokerage — natural-language property search over a live listings database.

Search · Qualification

Before you ask

The five questions everyone asks.

Which AI models can you integrate?

OpenAI GPT and Google Gemini by default, and Claude or an open model if you prefer. The workflow is model-agnostic — swapping providers later is a config change, not a rebuild.

Will it work with my CRM?

Yes, if it has an API or a Zapier/n8n node — HubSpot, GoHighLevel, Pipedrive, Zoho, Salesforce, or a Google Sheet if that's what you run on today.

Can it run on WhatsApp and my website?

Both, from the same brain. WhatsApp goes through the official Cloud API; the website gets an embeddable widget. Messenger and SMS work the same way.

Do I get the source code?

You get everything: the n8n workflow JSON, any Python services, prompts, environment template and deployment notes. It runs on your accounts, not ours.

How much of my time does this take?

Roughly an hour up front — one call to walk through your enquiry flow, plus access to your listings and calendar. After that you review a working build, not a spec document.

Next step

Send us your enquiry flow. We'll send back what it looks like automated.

Tell us where leads come in, what you ask them, and what your calendar looks like. You'll get a plan and a fixed price before anything gets built — usually same day.