AI agents for real estate: from first enquiry to booked viewing (2026).
What an AI agent can do for brokers, developers and property managers in the US, the UK and Dubai, what it should never do, and how to build one that buyers trust.
In real estate, the agent who replies first usually gets the viewing. Most enquiries arrive in the evening and at weekends, through portals, websites and WhatsApp, and many go unanswered for hours. An AI agent closes that gap. It replies in seconds, answers questions from your actual listings, qualifies the buyer or tenant, books the viewing in a negotiator’s calendar, and hands over with the whole conversation. We run one in production in real estate, on the same system that handles 70% of a solar company’s monthly sales.
This guide covers what an agent can do for brokers, developers and property managers, and the limits it needs.
What an AI agent does in real estate
Lead response and qualification
The agent replies to every enquiry from portals, your website or WhatsApp straight away, in the buyer’s language. It answers questions about the listing from your data (price, size, service charges, availability, nearby schools and transport) and asks the questions your negotiators would ask: budget, timeline, financing, cash or mortgage, first home or investment. Serious buyers get a viewing slot. Others get a follow-up plan.
Viewings and follow-up
It books viewings directly into the right person’s calendar, sends reminders and directions, and follows up afterwards. It also comes back weeks later when a similar property lists or a price changes, which is where many sales are actually won.
Off-plan and new developments
For developers, especially in markets like Dubai where off-plan sales dominate, the agent explains payment plans, handover dates and unit availability from live inventory. It handles the volume of questions a launch creates and passes qualified buyers to the sales team.
Property management
For lettings and management, it takes maintenance requests with photos and creates the work order. It answers routine tenant questions from the lease and building rules, and chases documents for renewals.
| Business | Highest-value job for the agent |
|---|---|
| Brokerage | Instant replies and viewing bookings for portal and WhatsApp leads |
| Developer | Off-plan questions, payment plans and launch-day volume |
| Property manager | Maintenance intake, tenant questions, renewals |
| Mortgage or conveyancing partner | Document collection and status updates |
What it should never do
An agent that talks to the public on your behalf needs firm limits, enforced in code rather than requested in a prompt:
- Never invent property facts. Every price, size and availability answer comes from your listing data. If the data doesn’t say, the agent says it will check.
- Never commit to terms beyond what you’ve approved: no price reductions, no promises about offers, no legal or tax advice.
- Never steer. In the US, the Fair Housing Act applies to automated conversations as much as human ones. The agent must not ask about or act on protected characteristics, and must offer the same information and options to everyone. Test for this explicitly.
- Respect consent for messages and calls. US rules on automated texts and calls require consent. In the UK and EU, data protection law governs follow-ups. Build opt-outs in from the start.
Buyers forgive an agent that says “let me check with the team.” They don’t forgive one that quoted the wrong price.
Why WhatsApp matters, especially in Dubai
In the UAE, much of the property market runs on WhatsApp. Buyers send screenshots of listings, voice notes and questions at midnight, and expect an answer. An agent there must handle Arabic, English and often other languages, and work within Meta’s rules for business AI. Our guide to building AI in the UAE covers the data rules, and Inside a WhatsApp sales agent shows how the conversation pipeline works.
What makes one work in production
- Clean listing data. The agent is only as accurate as your feed. Fix missing service charges and stale availability first.
- Calendar and CRM integration. An agent that can’t book or log is just a chatbot.
- A clear handover. Decide when a person takes over (a serious buyer, a complaint, a question it can’t answer) and make sure they get the full conversation.
- Testing on real conversations. Build the test set from your past enquiries, including the awkward ones, and score every change against it. See evals before features.
- Measurement. Track response time, viewing rate and conversion against the months before launch.
Questions property businesses ask
Will buyers know they’re talking to AI?
They should, and in the EU it’s a legal requirement for chatbots. In practice, buyers care more that the answer is fast and right. A clear introduction followed by accurate answers works better than pretending.
Can it work with portals like Zillow, Rightmove, Property Finder or Bayut?
Yes, through their lead notifications, email routing or the CRM they feed. The exact route depends on how each portal delivers leads to you.
Will it replace our negotiators?
No. It does the first reply, the qualifying questions, the scheduling and the follow-ups, so negotiators spend their time on viewings and closing.
Modulus Labs AI builds AI sales agents for real estate and other industries, on WhatsApp, web and email. Tell us how your leads arrive today, and we’ll show you what an agent would handle.