AI Real Estate Agent
Qualifies buyers on WhatsApp and books viewings that are genuinely free
My role: Designed and built end to end: matching engine, availability model, agent, dashboard.
The demo shown is the US configuration for a fictional agency, Lakeside Homes Realty in Austin, Texas. In-house product.
Overview
A WhatsApp agent for estate agencies. It captures what a buyer actually needs, searches only the agency's verified listings, explains why each match fits, and books viewing slots checked against agent hours, travel time and existing appointments.
The problem
Agencies get dozens of unqualified enquiries a day and lose good ones to slow replies. A chatbot that invents a listing, quotes the wrong price, or double-books an agent is worse than none.
The solution
The agent can only mention properties returned by a search in that conversation. Availability is computed from agency hours, agent calendars, blocked time and travel buffers, and two database constraints make it impossible for two viewings to take the same agent or property slot.
Principal features
Qualifies before it books
Captures budget, area, size and timeline, and keeps what the customer said separate from what the agent inferred.
Only verified listings
The model cannot invent a property or a price; it can only reference search results.
Slots that are really free
Agent hours, calendar blocks and travel time between viewings all feed the availability engine.
Fair-housing safe (US)
Matching uses objective criteria only. Neighbourhood, mortgage and lease questions go to a person.
Two markets, one engine
Dollars, square feet and virtual tours for the US; marla, kanal, lakh and crore for Pakistan, in English, Urdu or Roman Urdu.
Handover with the full picture
Staff receive the requirements, properties discussed, open questions and the last message, then hand back in one click.
Technical decisions
- Availability is a pure function over calendar data so it can be unit-tested against edge cases like back-to-back viewings across town.
- Overlap protection is enforced in PostgreSQL, not in the prompt, so simultaneous requests cannot double-book.
- Market-specific units, currencies and compliance rules are configuration, not forks of the codebase.
More AI & Automation
AI Receptionist
A 24/7 WhatsApp booking agent for appointment-based businesses
View project
AI Shopify Sales Agent
Selling and fulfilment inside WhatsApp, Instagram and Messenger
View project
Tailyr AI Concierge
A grounded, multi-store RAG assistant for high-value retail
View project