AI FRONT DESK HELPER
The first objection

Do customers want a person or an AI on the phone? What the surveys say, including the ones against us

The first thing a contractor says when I mention an AI receptionist is "my customers want to talk to a person." They're right, mostly, and I sell the AI. So here are the three surveys I've found from the last year, who paid for each one, what they actually asked, and what I think the honest conclusion is.

By Drew Koch, owner of Gladiators Painting and founder of AI Front Desk Helper · Updated October 13, 2026

Short version. Asked in the abstract, most people prefer a human on the phone: 85% in one survey of home-service callers, 68% in another across all industries. Asked a more realistic question, it tightens fast: offered an AI now or a human in 15 minutes, it's 56 to 44, and 58% say getting the problem solved matters more than who solves it. 53% of homeowners say they're fine with AI handling the first call. The one number that holds across every survey is 86% want to be told when it's an AI. And the comparison none of the surveys lead with is the real one for a contractor: nobody is choosing between a person and an AI. They're choosing between an AI and voicemail, because the owner is on a ladder. In the one survey that offered that choice, 78% picked the business where a human answered, 9% the AI, and 3% the one where nobody did.

Who's asking, and who paid

Every survey on this topic was commissioned by someone selling one side of it. That doesn't make the numbers wrong, but it decides which questions got asked and which got left out. So before the numbers, the sponsors:

SurveyWho they askedWhat the sponsor sells
ServiceForge, October 2025 (run by OnePoll)6,000 adults, US, Canada, UKHuman answering services for home-service companies
Housecall Pro, December 20251,000+ US homeownersField-service software, including an AI answering product
Lorikeet, April 2026 (Prolific panel)1,083 consumers, eight countriesAI customer-support software

One more thing to know before the numbers: only one of the three is about home-service customers end to end. Housecall Pro asked homeowners about contractors. ServiceForge asked the general public, but its headline question was a home-service one ("when calling a plumber or electrician"), while its other figures were about businesses in general. Lorikeet is about customer service across every industry and never mentions a trade; it's here for what it shows about how the question is asked, not as a home-service number.

I'm the fourth party with an interest: I built an AI receptionist inside my painting company and I sell it. Read everything below with that in mind.

The case against AI, as strong as it gets

ServiceForge's survey is the one to take seriously if you're skeptical, because it asked home-service callers specifically and it got the biggest sample. The numbers:

I believe all of that. Most of the AI people have met is a phone tree with better grammar, or a chatbot that can't do anything but apologize. If that's what "AI receptionist" means to a caller, 85% is low.

The case for, with the same honesty

Housecall Pro asked homeowners a simpler question and got 53% saying they're comfortable with AI handling their first call or chat. A majority, barely, from a company that sells the product. The same survey found 72% would pay up to 10% more for a contractor with a better customer-service reputation and 97% expect transparent pricing before hiring, which says something about what "service" means to a homeowner: being answered and told the truth, more than who does it.

Lorikeet's survey is the most useful because it asked the question three ways and the answer moved each time:

How the question was askedPrefer a humanPrefer AINo preference
In general, human or AI?68%7%25%
AI right now, or a human in 15 minutes?56%44%—
For something sensitive (money, health, security)?39%21%40%

And two more from the same study: 58% said whether the problem gets solved matters more than whether an AI or a person solved it, and half of the people who'd had a good AI experience said it improved their view of the brand. Acceptance tracked the stakes: 71% were fine with AI for a password reset, 29% for disputing a charge. Booking a painting estimate sits a lot closer to the password reset.

The question none of them led with

Every one of these surveys frames it as person versus AI. That's the choice a bank has. It isn't the choice a contractor has.

At my painting company, the person who would answer the phone is me, and I'm on a ladder. Before the AI, the choice my callers faced wasn't "Drew or a robot." It was "voicemail or the next painter on the list." I counted the calls: 86% of the real ones came on weekdays between 8 and 5, while I was working.

ServiceForge did ask this one, buried in the middle of their report, and it's the most useful number in any of the three. Given three businesses with similar reviews, callers picked:

Read that the way a one-truck contractor should. If you can staff a phone with a good person all day and after hours, do it; 78 beats 9 and it isn't close. If you can't, and most of us can't, the AI beats the empty line three to one. The 29% who hang up on an AI were going to hang up on your voicemail too, and the voicemail doesn't book the other 71%.

What the numbers say to do, if you use an AI

Taken together, the three surveys are close to a checklist. I built ours to it, and I'd hold any product to it.

  1. Tell people. 86% want to know. Ours says it's an AI assistant whenever anyone asks, on calls and in texts, and never claims to be a person. There's a switch to say it in the greeting too, for companies and states that want that. Pretending is the one thing that turns the 85% into a lost customer instead of a booked one.
  2. Hand off the moment they ask for a person. 83% have asked. The wrong answer is "I'm an AI but I can still help, what do you need?" The right answer is "Of course," and a transfer, or a message if there's nobody to transfer to. We learned that one from a real call and fixed it the same day.
  3. Resolve, don't deflect. 58% care about the outcome more than the agent, and only 24% say AI usually resolves things. For a contractor, "resolved" means a time on the calendar and a confirmation text, not a promise to call back. If the AI can't book, it's a chatbot, and the 85% are right about it.
  4. Keep the stakes where AI belongs. Booking an estimate, answering hours and service area, giving a ballpark from your own price sheet: fine. Pricing a custom job, settling a dispute, promising insurance coverage: no. The surveys say acceptance drops as the stakes rise, and the product should too.
  5. Measure it on your own line. Every number above is someone else's customers. Yours will tell you in your transcripts and your booking rate. Ours reviews every call Alex takes against a checklist, so I can see when a caller wanted a person and whether she handed them off.

What I'd say to the contractor who asked

Your customers do want a person. So do mine. The survey that says 85% is probably right. But that's not the decision you're making. You're deciding what happens to the 9 a.m. call while you're cutting in a ceiling, and the choices are an AI that answers, books and tells the truth, or a voicemail that does none of it. On the only survey question that offered that choice, the AI won three to one.

If you've got a good person on the phone all day, keep them, and put the AI behind them for lunch, the other line and after hours. If you haven't, the AI is the person you don't have. Either way, let it say what it is.

Questions contractors ask

Do customers prefer talking to a human over an AI?

Asked in the abstract, yes: 85% of home-service callers in ServiceForge's October 2025 survey and 68% of consumers in Lorikeet's April 2026 survey. Asked whether they'd take an AI now over a human in 15 minutes, it tightens to 56–44, and 58% say resolution matters more than who provides it.

What percentage of people hang up on an AI receptionist?

29% say they hang up when they reach an AI instead of a person, in ServiceForge's survey. The same survey found only 3% would choose a business where nobody answered, against 9% for one where an AI answered.

Are homeowners comfortable with AI answering a contractor's phone?

Housecall Pro's December 2025 survey of 1,000-plus US homeowners found 53% comfortable with AI handling the first call or chat. Housecall Pro sells an AI answering product, so read it with that in mind.

Should an AI receptionist say it's an AI?

Yes. 86% of people want it clearly indicated. Ours says so whenever a caller asks, never claims to be human, and has a switch to say it in the greeting as well.

What should happen when a caller asks for a real person?

A transfer, right away, with no sales pitch for the AI. If there's nobody to transfer to, a message with a promise of a callback. 83% of people have asked for a human at least once, so this is the most common thing the AI will be asked to do.

Is an AI receptionist better than voicemail?

On the one survey question that compared them, 78% chose a business where a human answered, 9% one where an AI answered, and 3% one where nobody answered. If you can't staff the phone, that's the comparison that matters.

Where do these numbers come from, and are they about home services?

ServiceForge (sells human answering services) via OnePoll, October 2025, 6,000 adults in the general public, with its headline question about calling a plumber or electrician; Housecall Pro (sells field-service software with an AI product), December 2025, 1,000+ US homeowners asked about contractors, the only one that is home-service end to end; Lorikeet (sells AI support software), April 2026, 1,083 consumers in eight countries about customer service in general. Each is linked on this page.

See it on your own phone line

Fifteen minutes. We call your business number, run a real call, and show you what happens after it. You decide.

Book a demo

Related: why an AI receptionist for a home service business, answering service vs AI receptionist, what callers ask an AI receptionist, and how contractors lose jobs to voicemail.