What to look for in an AI answering service for landlords
Short answer: a good AI answering service for landlords picks up every call, says plainly that it’s an AI, checks who is calling, asks the right questions, tells the tenant what happens next, and brings you a decision. It should not pretend to be a person, make your decisions, spend your money or open anyone’s door on its own. Here is a checklist of both lists, plus questions to ask before you hand over your tenant line.
Keystead Property Office. Published .
I’m a landlord in the Piedmont Triad. I built one of these, Keystead, an AI receptionist for landlords who self-manage, so I’ll be upfront about that. I’ve tried to write this so it’s useful whichever service you choose, including none.
General information, not legal advice. This article mentions the FCC’s ruling on AI-generated voices and North Carolina’s repair statute, G.S. 42-42. Rules change; ask an attorney how they apply to your setup.
What an AI answering service is
It’s software that answers the phone (and often texts) for your rentals. It uses a synthetic voice and a language model to hold a conversation. It’s different from a voicemail box, which just records, and from a human answering service, which has people reading a script. A landlord-focused one knows who your tenants are, which unit they live in, how to sort a leak from a lost key, and what to do next.
What it should do
1. Say it’s an AI, in the first sentence
The call should open by identifying itself as an AI assistant, plus a notice that the call may be recorded. It should answer honestly if someone asks “are you a person?” and let them reach a human on request.
This isn’t just manners. AI voice rules are tightening. The FCC has ruled that calls using AI-generated voices are “artificial” under the TCPA (FCC announcement), and a service that hides what it is invites trouble.
2. Answer every call, any hour
That’s the whole point. Ask what happens at 2 a.m., on a holiday, and when five tenants call at once. Ask what happens when the AI can’t understand someone. There should always be a path to a person.
3. Check who’s calling
Before it talks about a tenant’s home, the service should confirm the caller is who they say. Matching the phone number to a tenant record is a first step; asking for a name and address is a second. It matters most before anything involves access. A service that discusses your tenants’ details with any caller is a privacy problem.
4. Triage, and send true emergencies straight to you
For a gas smell, a carbon monoxide alarm or anyone in danger, it should tell the caller to get out and call 911 first. For a no-heat call in freezing weather, a burst pipe or a lock that won’t secure the home, it should tell you right away and keep escalating until someone acknowledges. We lay out a sample triage in what counts as a rental emergency in North Carolina. The service should follow rules you set, and show you what it decided and why.
5. Ask clear questions, one at a time
What’s wrong, where, since when, is it getting worse, when is someone home, may a repair person enter, are there pets. A tenant on a stressful call shouldn’t get a barrage. A human-sounding pace matters more than speed.
6. Be honest about what happens next
“I can’t promise a time, but I’ll check with the plumber and text you a confirmation” is the right kind of sentence. A service that promises “someone will be there in 30 minutes” when it can’t know is setting the tenant up to be angry at you.
7. Confirm in writing
A text with a reference number, then a confirmation of the scheduled window, a reminder, and an “on the way” note. That’s what turns a phone call into a record.
8. Keep a log
Every call, text, approval and entry should be time-stamped and findable. If there’s ever a question about when a tenant first reported something, the log answers it.
What it shouldn’t do
1. Pretend to be a human, or hide that it’s AI
No fake first names with a “just a second, let me check” pretending to type. If a tenant asks “are you real?” the answer is “I’m an AI assistant,” followed by an offer to get a person.
2. Dispatch, spend or unlock without your approval
It can gather the facts and propose: “Send this plumber, up to $300?” You tap approve. Letting a vendor into a tenant’s home, giving out a door code, and spending your money are your decisions. If a service lets you pre-approve small jobs, that should be a setting you control, not a default.
3. Give legal advice or make legal decisions
It shouldn’t tell a tenant what their rights are, whether you’re liable, whether a lease is broken, or how an eviction works. When a call drifts into legal territory, accommodation requests, threats or distress, it should escalate to you, not improvise.
4. Deny anything on its own
Accommodation requests, habitability complaints and anything touching safety should go to you immediately. An AI that says “no” to a tenant on your behalf is the kind of thing that ends up in a letter.
5. Text or call people who haven’t agreed
If a service sends automated texts or places calls, ask how it records consent, honors STOP, and respects quiet hours. Tenants are people in their homes; a text at 2 a.m. about something that isn’t urgent is a bad idea. The standard for marketing messages is higher than for service ones, and a service shouldn’t use tenant numbers for marketing at all.
6. Sell or share your tenants’ data
Ask what it stores, for how long (call recordings especially), who can see it, and whether it’s ever used for anything other than your account. You should be able to get your records out.
Questions to ask any provider
Copy this list:
- What does the call say in the first sentence? Does it say it’s an AI and that it may record?
- What happens when the AI can’t help? How does a tenant reach a person?
- How does it verify callers before sharing details or arranging access?
- Which situations reach me immediately, and how do I know I got the alert?
- Does anything get dispatched, spent or unlocked without my approval?
- How does it record tenant and vendor consent for calls and texts? Does STOP work immediately?
- How long are recordings kept, and who can hear them?
- Can I see a log of every call, text and visit?
- What does it cost per door, per month, and what is the minimum? Are there long contracts?
- Can I try it first, with real numbers I can check?
Any honest provider will answer these without hedging. If one dodges a question, that’s information.
AI answering service or human answering service: the honest case
For: the phone is the part of self-managing that never stops. If you have a day job and several doors, the same person (you) takes every call and writes everything down afterwards. Handing the first conversation to something that never sleeps, and asking it to bring you decisions rather than interruptions, can give you real time back.
Against: it costs money. With one to three doors and rare after-hours calls, a notes app and a good emergency card (see after-hours maintenance calls: a plan for landlords who self-manage) will do. Some tenants don’t love talking to an AI; the answer is a clear path to a person and an honest greeting. And no AI will make a good decision about a messy situation on its own; you stay in the loop.
A simple test: add up the after-hours calls and texts you handled last month, and the time each took, including the follow-up. If it’s under about three hours, you may not need a service. If it’s much more, or it’s costing you sleep, a trial is worth a look.
Where Keystead fits
Keystead is an AI receptionist and maintenance line for landlords who self-manage, built for roughly four to twenty doors. Juno, the AI assistant, checks the caller, triages, texts a reference number, asks the vendor for the soonest time, and brings you one decision. Asked whether it is a person, it says plainly that it is an AI. You approve every dispatch and every code. It’s software you control, it doesn’t collect rent or hold money, and it isn’t a person who takes over your rentals.
Keystead is a pilot today for landlords in the Triad, and I want honest feedback from the people who use it. If it isn’t a fit, I’ll tell you, and I’m happy to point you to a free tool that is.
Frequently asked questions
Does an AI answering service have to tell callers it’s an AI?
It should never hide it, and some states require it up front. Keystead says it is an AI whenever asked, and in the first sentence for callers from states that require it. Rules on AI voices and call recording are tightening, and tenants deserve to know who they’re talking to. The FCC has also ruled that AI-generated voices count as artificial voices under the TCPA, which affects calls and texts that need consent. Ask your attorney how it applies to your setup.
Can an AI answering service dispatch a plumber without my approval?
It shouldn’t. A good one gathers details and brings you a clear choice, such as “send this plumber, up to $300?” Spending your money and letting people into a tenant’s home are your decisions. Some services let you pre-approve small repairs; that should be your setting, not theirs.
Is an AI answering service worth it for a few rental units?
It depends on how many after-hours calls you get and how much they cost you in time and sleep. With one to three doors and few calls, probably not. With a handful of doors and a day job, the phone can become the biggest time cost. A trial, with real numbers, will tell you.
Sources
- FCC, AI-generated voices in robocalls are illegal without consent (TCPA ruling)
- FCC, Stop unwanted robocalls and texts
- FTC, CAN-SPAM Act: a compliance guide for business
- NC General Assembly, G.S. 42-42
General information, not legal advice.
Related
See it on your own tenant line
Keystead is a pilot for landlords who self-manage. Tell us about your rentals and we will show you how it works.