Voice AI or a Live Desk? 5 Home Service Cases, With Numbers

The short version: In home services, coverage is the product, and the examples below show the stake. A plumbing franchise closed 62% of after-hours calls into jobs, and an HVAC company booked $477,908 in new revenue. Some of these businesses were running an AI system; some were paying people to answer. Each case says which.

Most shops measure the calls they answer. The call that never becomes a job is the one that eats the week.

A homeowner with water on the floor or a house with no cooling has a short list of numbers to try, and they are not browsing. The company that picks up and books the work gets paid.

Five published examples follow. Two are live answering desks in Baton Rouge, one is a nationwide mover that could not call back people who hung up, one is a North American operator with a network of local brands, and one is a single-site repair shop in Southern Florida. None of them are clients of this site. A wider set sits in a roundup of real voice and answering results.

Does voice AI actually book jobs in home services?

Yes, when it is pointed at the calls a small team cannot reach. Across the five published cases below, coverage shows up as closed jobs, booked revenue, and callers who hung up and came back. Two of the examples used people rather than AI, and each one is labelled so you know which.

Coverage is how the job starts in this trade. A sewage backup does not wait for a morning callback, and the caller dials the next listing while yours rings out. Your price was probably fine. Nobody was there to hear it. A close rate, a booked-revenue total, and a week-one payback measure different gaps, so copy the one that matches the leak you actually have.

What an after hours answering service did for a plumbing franchise

Situation

A single-location plumbing franchise in Baton Rouge, Louisiana. After-hours calls were the gap. Relaying the message accurately, and booking the job, mattered more than a polite greeting.

Someone with water on the floor is working down a list. If a person answers and books a tech, that company owns the job. If the call hits voicemail, the homeowner dials the next name.

What changed

The company implemented 24/7 live-voice answering, customized scripting, and an enhanced ServiceTitan integration. ServiceTitan is the scheduling software. The integration kept the night call out of a notepad someone would retype in the morning. The same kind of answering setup also integrates with Housecall Pro. This franchise used the ServiceTitan side, so the booking landed where the schedule already lived.

A plumbing answering service is only as good as the handoff. A wrong address, or a vague "they have a leak," wastes a truck. The script tells the person who answers what to ask and when the job is real enough to book.

This is live-voice answering coverage, with people taking the calls rather than an AI voice agent.

Billing is per minute. Holds where the caller waits for service are never billed, and only the minutes where someone is actively working the call are. If you are comparing answering service cost, no rate is published for this case, so there is no figure to quote here. The principle still holds: you should not pay for the time a caller spends waiting.

Results

62% of after-hours calls closed into actual jobs.

What the published account reports	Detail
After-hours calls closed into jobs	62%
Coverage	24/7 live-voice answering
Scripting	Customized for the franchise
Scheduling	Enhanced ServiceTitan integration
Related software	Also integrates with Housecall Pro
Billing	Per minute; waiting holds are not billed

The 62% is the share of after-hours calls that became jobs. It is not a revenue total, and no before figure is published. A person, a script, and a ServiceTitan booking turned a majority of those night calls into work on the board.

What a Baton Rouge HVAC company booked

Situation

An HVAC and electrical company in Baton Rouge, Louisiana. Two trades, one phone line. A no-cool call does not wait for Monday dispatch, and an electrical call after dinner is not a morning problem. No miss rate is published, so none is invented. The published story is what happened after a person was on that line all the time.

What changed

The company implemented 24/7 live answering in front of the phones, a ServiceTitan integration, and customized call scripting. People answered. An HVAC answering service fails if the script cannot split a cooling call from a panel call. Those are not the same dispatch.

A 24/7 answering service, in this case, meant someone picked up when the office could not, asked what the office would ask, and wrote the job where the dispatcher already lives. A conversation that dies in a text to the owner is not a booked job.

Management at the company described the relationship this way: "The dedication Dexcomm has to ensure my satisfaction is above and beyond what I've ever experienced from a vendor. They truly care about their customers and are willing to do whatever it takes to ensure satisfaction."

That is a satisfaction quote, not a revenue quote, and the dollar figure belongs to the published case study rather than to the testimonial. Like the plumbing franchise, this was live answering coverage with people taking the calls.

Results

$477,908 in new booked revenue, plus an increase in booked jobs and revenue.

What the published account reports	Detail
New booked revenue	$477,908
Booked jobs	Increased
Revenue	Increased
Coverage	24/7 live answering
Scheduling	ServiceTitan integration
Call handling	Customized scripting

The $477,908 is vendor-reported new booked revenue, not profit, and not an independent audit. Jobs on the board are not cash in the account. Do not paste the figure into your forecast and call it a plan.

A nationwide mover and a 64% call abandonment rate

Situation

A large independent moving company in the United States, about 20,000 moves a year, working nationwide. Someone calling a mover usually has a date, an address, and a budget. They are close to a decision.

The legacy phone system could not call back people who hung up. If a customer left before reaching a person, or chose not to leave a voicemail, there was no record and no follow-up. That was a 64% call abandonment rate. Nearly two out of three callers hung up without a conversation, and often without a trace. You lost them to a phone system that could not see them leave.

What changed

The company implemented automatic callbacks, automated voicemail flows, and a move of scheduling off spreadsheets into workforce management. If someone hung up before reaching an agent, the system called them back. Scheduling left the spreadsheet so the right number of people were on the phones when volume showed up.

This is an AI-assisted contact-center platform: software handled routing, callbacks, and the flow around the call, while people still answered. It was not a voice agent, and it was not a live answering desk in an otherwise empty office.

Results

Metric	Published result
First-call resolution	up 41%
Supervisor escalations	down 30%
Abandonment before the change	64%
Revenue	up 74%
New revenue stream identified	$1M in 30 days
Process adherence	up 48%
On-time pickup / delivery	100% / 98%

Read the 74% with a cool head. A company that is also adding people can grow for more than one reason, so treat it as the published business outcome rather than a receipt for callbacks alone. The $1M figure is a new revenue stream identified in 30 days, and identified is the verb. It is not cash collected. The callback itself is easy to copy at any size: see a missed call, ring it back the same day.

Reviews and revenue at a multi-brand operator

Situation

A North American home services operator in HVAC, electrical, and plumbing, with a network of local brands. No city is named in the facts used here, so none is invented. Nationwide would be the wrong word. North American, and multi-brand, is what the account supports.

What changed

The published account is about data and AI modernizing how the operation runs. It is not a write-up of a voice agent on the phones. No script, no after-hours close rate, and no ServiceTitan detail belong here. Those sit with the Baton Rouge companies.

Results

Combined, the local brands average 4.9/5 on Google reviews, and residential HVAC revenue grew more than 30% year over year.

What the published account reports	Detail
Google reviews, combined local brands	4.9/5
Residential HVAC revenue	up more than 30% year over year
Trades	HVAC, electrical, plumbing
Footprint	North American network of local brands

A 4.9 review average is a reputation number rather than a call-answer rate, and growth above 30% year over year is a business outcome rather than a phone receipt. This account is about data and AI inside the operation, so copy the scale you actually have.

A Southern Florida shop that covered a year in a week

Situation

A single-site automotive repair shop in Southern Florida, specializing in European engine control units. Leads arrived on the web and by text, including while the bay was full. The published change is not a voice agent on the main line.

What changed

The shop added 24/7 live chat plus a virtual receptionist to handle web and SMS leads, so an inquiry got a response instead of sitting in a form until Monday. The channel here was chat and a receptionist rather than a voice agent, and a chat result is not a phone result even when the payback is fast.

Results

In the first week, 20 new leads. 3 of them bought $1,000 products. That paid for an entire year of service.

What the published account reports	Detail
New leads in the first week	20
Buyers of $1,000 products	3
Payback	A full year of service, in that week
Channel	24/7 live chat and a virtual receptionist
Footprint	Southern Florida, about a single site

Smith.ai's published account puts the average cost per lead in the automotive sector at $205. That is a sector average, not this shop's cost, and it is not turned into a savings claim the case study did not make. Twenty leads in a week is a real pile of demand for a single site. Three bought, and the service paid for a year. The channel was chat and SMS.

Photo by Fiqih Alfarish on Unsplash.

What to copy first

An unanswered inquiry leaves, whether it is a night call, a caller who hung up, or a web lead sitting in a form. What turns it into a job is the handoff. Scripting and ServiceTitan both showed up in Baton Rouge because the person answering is often not the technician, and a note that never reaches the schedule is a conversation you pay for twice.

It is worth keeping the labels straight even if you only care which one to buy. The live answering in Baton Rouge is what the 62% and the $477,908 belong to. The mover ran an AI-assisted platform while people still took the calls. The multi-brand result came from data and AI inside the operation, not from a voice agent. The shop used chat and SMS. If a pitch uses one of these numbers to sell you a different product, it is worth walking away.

If you are pricing an answering service for contractors, three questions sort most of the field. Can the person who answers book the job, or only take a message? Will the booking land in ServiceTitan or Housecall Pro if that is what you already run? And are you billed for the minutes a caller spends waiting? In the plumbing example, waiting holds are not billed. No rate is published there, so price belongs in the quote in front of you.

Copy in this order:

1. Cover the hours you are missing. A same-day callback is free, and a live after-hours desk is what the Baton Rouge companies bought.
2. Call back people who hang up. The mover's leak was abandonment with no record of the caller, and ringing the missed call back the same day costs nothing.
3. Write the script before you outsource the voice. It needs the symptom, the address, whether water is spreading or the house has no cooling, and an arrival window you can actually offer.
4. Book into ServiceTitan or Housecall Pro if that is what you dispatch from. A text thread nobody checks at 7am is not a schedule.
5. Measure the leak. Answer rate flatters you, while closed jobs, booked revenue, and abandonment are the numbers these examples published.

After that, the question is whether you keep paying for live coverage or build a system you run yourself. Either way the shape of the fix is always-on intake for home services: answer the call, capture the details, book the work, and relay it cleanly into the schedule.

If you want to build this

The first version is a phone and a person who will answer it. Do that before you buy anything. When the missed-call list gets long enough to automate, you have two paths, and they are not the same product.

Keep a live desk, and wire the handoff so nothing sits in a notebook. Make.com can watch for a missed call or a completed booking and push it into the system your dispatcher already opens.

Or have software answer when nobody is on shift. Voice-agent platforms like Retell AI and Vapi are where most teams start. That path is AI. It is not what the Baton Rouge companies ran. The mover's platform assisted the people who still took the calls. A voice agent is a further step.

If you go the agent route, the voice decides whether the caller stays on the line. A November 2025 Twilio blind test found 90% of consumers could not correctly identify an AI voice from a human one, though 72% believed they could tell. Neural voices from ElevenLabs are a large part of why that gap exists. You still need the script, the handoff, and a rule for when a person takes over.

The AI version of this front desk is written up in how an AI receptionist works for home services. Same leak. Different tool. Label it before you buy it.

Not sure which layer your phones need? Book a free discovery call and we will map the calls you are missing to the jobs they should have been, without pretending a live desk and a voice agent are the same thing.

The numbers in this article come from each provider's published case study: Dexcomm for the Baton Rouge plumbing franchise and the Baton Rouge HVAC and electrical company, Talkdesk's customer story and Observe.ai's customer spotlight for the nationwide mover, Lumenalta for the North American home services operator, and Smith.ai for the Southern Florida automotive repair shop. These are real anonymized case studies that show what voice AI can do in practice. Some links in this post are affiliate links: if you sign up through them, we may earn a commission at no extra cost to you.