AI answering service for lettings agencies: what it should and shouldn't handle

The short answer

A lettings agency's phone rings for things that do not keep office hours: a boiler that fails on a Sunday, an applicant who can only call after work, a landlord abroad in another time zone. An office open 9:00 to 17:30 on weekdays is closed for about three quarters of the week, and the calls that arrive then mostly go to voicemail.

An AI answering service is useful here for a narrow, specific reason: it can complete the routine calls and correctly hand on the rest. It should never be asked to make the decisions a lettings agent is paid to make. This article is about that split, and how to set it up so nothing dangerous falls through it.

The coverage gap, in arithmetic

Take an office open 9:00 to 17:30, Monday to Friday: 8.5 hours × 5 days = 42.5 hours, out of the 168 in a week.

That is 25.3% of the week covered and 74.7% not. Add a Saturday morning and the picture barely moves. The same arithmetic applies to a single receptionist, which we work through in AI receptionist versus human.

What arrives when the office is closed

Every agency's mix is different, so count yours. As an illustration, take a week with 30 out-of-hours calls: 12 viewing requests, 10 repair reports, 4 applicant questions, 2 rent or account queries and 2 genuine emergencies.

In that illustration, an AI service could complete 16 of the 30 calls without a person — booking viewings into real availability and answering applicant questions from your own information. It would log 12 properly for the morning, with the details a contractor or account handler needs, and escalate the 2 emergencies at once. Nothing is left on voicemail, and nobody's evening is spent listening to messages.

Emergencies come first, and are never handled by the AI

The design decision that matters most is not what the AI can do, but what it must hand over immediately. In our view, an out-of-hours service for a lettings agency should escalate to a named on-call person, at once, for:

• A smell of gas or a suspected carbon monoxide leak — telling the caller straight away to call the National Gas Emergency Service on 0800 111 999 (free, 24 hours) and follow its advice.
• Flooding, a serious leak, fire or electrical danger.
• A property that is not secure, such as a break-in or a door that will not lock.
• No heating or hot water where someone vulnerable lives.
• Anyone distressed, at risk, or raising a safeguarding concern.

Write these rules down before anything goes live, and test them with real scenarios. A service without a tested escalation route is not finished. We cover the general questions to ask any provider in AI answering services for UK businesses.

Where AI is heading in property

Property businesses are adopting AI at about the national average. In June 2026, the Office for National Statistics found 25.1% of real estate businesses using at least one AI technology — and most of them only one.

That leaves room for an agency to get ahead on service, provided the AI sits on top of good systems. Answering the phone is only valuable if the viewing lands in the real diary, the repair lands in the maintenance system and the landlord sees the same record — which is the subject of the systems a lettings agency runs on.

Setting it up well

1. Count two ordinary weeks of out-of-hours calls by type before buying anything.
2. Write the escalation rules first, with names and phone numbers, then the routine scripts.
3. Connect it to your real systems — the viewing diary, the maintenance log, the tenancy record — so calls become records, not messages. See AI calling agents.
4. Tell callers they are speaking to an AI, plainly, at the start.
5. Review a sample of calls every week for the first month, and adjust.

For how we approach whole-agency platforms, see software for real estate and lettings.