How Estate Agents Use AI Automation to Answer Property Enquiries

The short version: AI automation reads an incoming property enquiry from a portal, email or WhatsApp, matches it to the right listing in your CRM, sends a personalised reply within a minute or two, offers viewing slots, and only hands the conversation to a human when the buyer or tenant is ready to talk terms. Done well, it turns a Friday night enquiry that used to sit unread until Monday into a booked Saturday viewing. Done badly, it's just a chatbot that says "thanks for your interest" and annoys everyone.

What "AI automation" looks like on a Tuesday afternoon

Forget the demo videos. Here's what it looks like in a real branch. Someone submits an enquiry through the Rightmove contact form for a two-bed flat in your window. Within 90 seconds they get a reply that mentions the flat by name, the street it's on, the council tax band, whether pets are allowed, and three available viewing times this week. If they reply "yes to Thursday at 5", it books straight into the negotiator's calendar without anyone touching a keyboard. If they ask something the system doesn't have an answer for, like whether the freeholder allows a home office extension, it says so honestly and flags a human to call.

That's the whole job. Not "engagement", not "customer experience". Answering the question, fast, with the right facts, and getting a viewing in the diary. Everything else is decoration.

The uncomfortable bit about response speed

Most people writing about this will tell you speed is everything. It matters, but not for the reason most agents assume. A well known lead response study found that contacting a lead within five minutes makes you roughly 21 times more likely to qualify them than waiting thirty minutes. Agents read that and think "great, AI will win me more deals." What it does is stop you losing deals you'd already half won. Property managers and letting agents with a process worth sharing can write a property management guest post.

Here's the part nobody wants to say out loud: most enquiries that sit unanswered for six hours were never lost because AI wasn't clever enough to reply. They were lost because the branch had one person covering three portals, a phone that never stopped, and a policy of "we'll get to it after viewings." AI automation doesn't manufacture new buyers. It stops you leaking the ones you already paid Rightmove and Zoopla portal fees to attract. That's a smaller, less exciting story than the sales pitch, but it's the true one, and it's still worth thousands of pounds a month to a busy branch.

How the enquiry flows, step by step

This is roughly how a working setup runs, whether the agency has built it themselves or brought in outside help:

• An enquiry lands from Rightmove, Zoopla, OnTheMarket, the agency's own website form, or WhatsApp.
• The AI system pulls the listing details from the CRM (Reapit, Jupix, Alto and Street are the common ones in the UK) so the reply is specific, not generic.
• A first reply goes out within one to two minutes, answering the obvious questions: availability, price, whether it's still on, key features, and next steps.
• The system offers real viewing slots pulled from the negotiator's actual calendar, not a made-up list, so double bookings don't happen.
• If the person books, it confirms by text and email and adds a reminder 24 hours before. This is the same logic covered in more detail in this piece on how real estate agents can use AI for appointment scheduling, and it's usually the single biggest time saver in the whole setup.
• If they don't reply within a day, a soft follow up goes out on day one, day three and day seven, then the lead is marked cold and moved off the human's priority list.
• Anything that smells like a serious buyer or a complaint gets flagged straight to a person, no automation attempts to close it.

A real example: the Essex agency and 140 enquiries a week

I worked with a lettings and sales agency in Essex last year that was getting around 140 enquiries a week across three portals and their own site, split across four branches. One negotiator was handling first response for the whole patch because the branch managers were busy doing viewings and valuations. Their own numbers showed the average first reply took 6 hours during the week and over 20 hours if the enquiry came in on a Saturday afternoon, because nobody checked the inbox again until Monday.

We set up an AI system to handle first response, viewing booking, and the basic FAQ (deposit amount, move-in date, pet policy, whether bills were included). Within the first month, average first response dropped to under two minutes. Viewings booked without a human touching them went from zero to about 35% of all bookings. The negotiator who used to spend her mornings replying to "is this still available" messages was freed up to call the warm leads, the ones asking about mortgage in principle or move-in dates, which is where her time was worth something.

The branch manager told me the thing that surprised her most wasn't the speed, it was how many enquiries turned out to be dead the moment someone replied. People who'd enquired on five flats at once and had already found somewhere else. Automation didn't create those tyre kickers, it just found them faster so the negotiator stopped chasing them for a week. That's the same pattern I've seen written up elsewhere with sharper detail in this piece on never letting a real estate lead go cold, and it holds true across every agency I've looked at this closely.

What it costs and what it's worth

Rough UK pricing as of 2026: a basic AI enquiry handling setup built on an existing CRM integration runs somewhere between £150 and £400 a month for a single branch, depending on enquiry volume and how many portals and channels it covers. A multi-branch agency building something more custom, with WhatsApp, voice call handling and CRM write-back, is looking at a proper build cost upfront (often £2,000 to £6,000 depending on complexity) plus a smaller monthly fee for hosting and support. There's a full breakdown of how those numbers get built up on the page about what an AI consultant costs, which is worth reading before anyone quotes you a number that sounds too neat.

Worth it against what, though? Portal fees alone often run £800 to £1,500 a month per branch for Rightmove and Zoopla combined. If automation recovers even one extra instructed property a month that would otherwise have gone to the agent who replied faster, it's paid for itself several times over. The maths is boring but it's the maths that matters, not the demo.

Where AI should stop and a person should take over

This is the part agencies get wrong most often, usually by trying to make the AI do too much. It should never be the one negotiating price, handling a complaint about a damp survey, or explaining why a landlord won't accept a guarantor. It shouldn't pretend to be a person if someone directly asks whether they're talking to a bot, that's a fast way to lose trust with someone who's about to sign a twelve month tenancy. And it shouldn't be left completely unsupervised, because property details change (a flat goes sold subject to contract, a landlord pulls out) and an AI system replying with outdated availability makes the whole agency look sloppy.

The clearest way to think about it is the same framework I use for any business function, laid out in when you should use AI agents in your business, and when you really shouldn't: automate the repeatable, factual, high volume first step, and keep a human on anything that involves judgement, money, or an upset person. Property enquiries are almost entirely the first kind. Complaints and negotiations are almost entirely the second. Agencies that blur that line are the ones who end up with a chatbot arguing with a tenant about a broken boiler at 11pm.

If you want to see how this plays out in other parts of the business, not just enquiries, there's a good roundup of real-world examples of AI agents that are working in 2026, and a fuller look at the admin side specifically for this sector in what gets fixed in letting and sales admin, which covers referencing, compliance chasing and the paperwork nobody enjoys, alongside the enquiry handling covered here.

The bit that still needs a human voice

None of this replaces a negotiator who knows the street, knows the landlord, and can read a buyer's tone on the phone in a way software can't. What it does is get rid of the boring, repetitive first mile of every enquiry so that human time gets spent where it changes an outcome. The agencies doing this well aren't the ones with the flashiest chatbot. They're the ones who worked out exactly which three questions get asked a hundred times a week, automated those, and left everything else to the people who are good at their jobs.

This guide is part of my AI Consultant by Industry: What to Automate First in 19 Sectors.

Related: ai consultant uk estate agents what to automate first.

Related: ai consultant contract what should be in it for small businesses.

Related: automation (this site).

If you're a broker who works with estate agents, here's how to pitch a mortgage article for my blog.

Frequently asked questions

Do sellers and landlords know when AI is handling the first reply?

Most agencies don't advertise it loudly but don't hide it either. Good practice is to disclose it if directly asked, and to make sure the AI is clearly speaking as "the team at [agency name]" rather than pretending to be a named negotiator who doesn't exist.

Will an AI system book a viewing straight into an agent's calendar without checking?

Yes, if it's set up it reads real availability from the negotiator's calendar or the branch's booking system, so it can't double book. This is the same logic used for appointment scheduling automation and it's usually the fastest win in the whole setup.

What happens to enquiries about properties that have already gone under offer?

A well built system pulls live status from the CRM before replying, so it tells the enquirer the property is under offer and offers similar listings instead, rather than sending a stale reply. This needs the CRM and the AI tool to be connected, which is where a lot of cheap, off-the-shelf chatbots fall down.

Is this only worth it for large agencies with high enquiry volume?

No. A single busy branch getting 30 to 50 enquiries a week can still see a real return, mainly through weekend and evening cover, since most missed enquiries happen outside office hours when nobody is checking a portal inbox at all.

Further reading

• Anthropic's own documentation
• the NIST AI Risk Management Framework

If you would rather pitch your own version, you can write for us about home and interiors.