Blog / AI receptionists

What missed calls cost a service business, and what an AI receptionist actually does about it

In the largest recent study of business calls, 44% went unanswered and home services missed 48%. Here is what that data can and can't tell you, and what an AI receptionist does about it.

In the largest recent study of business phone calls, 44% of inbound calls went unanswered. Home services missed 48%. Senior care missed 47% and healthcare 46%.

If you run a service business, some of those calls were yours. This post covers what the research says a missed call costs, where the numbers are shaky, and what an AI receptionist does about it. It also covers what an AI receptionist does not do, which matters just as much.

What the call data says

Those figures come from Invoca’s 2026 analysis of 70 million inbound calls across 10 industries, published in July 2026. We are working from JustCall’s summary of the Invoca report, not the full report, so treat the exact numbers with some care.

The same summary has two more figures worth knowing. Of the calls that were answered, 38% were qualified leads, and 42% of those qualified callers converted on the call. Inside home services, answer rates ranged from 32% to 74% depending on the trade, so the average hides a lot.

Where the numbers disagree

Not every dataset agrees. A CallRail benchmark of small businesses reportedly put the home-services miss rate at 14%, according to PipelineOn and Stealth Agents. Studies define a “missed” call differently. Some count any call no person picked up. Others count only callers who hung up.

Two figures you will hear in sales pitches are worth ignoring. The claim that 62% of calls go unanswered comes from a 2016 study of 85 businesses. The claim that a business loses $126,000 a year to missed calls has no traceable source at all. JustCall went looking and found nothing behind it.

The fair reading is this. Many service businesses miss a large share of their calls, the share varies widely, and your own phone records are the only number that really counts.

What happens after the call goes to voicemail

Callers don’t wait long. In a CallRail survey of 1,000 consumers from September 2025, reported by JustCall, only 42% said they would leave a voicemail. Another 21% said they would immediately call another business. That is a vendor claim, and it measures what people say they would do, not what they were seen doing.

Slow follow-up makes it worse. In a secret-shopper study of 1,333 US law firms in early 2025, Hennessey Digital found a median of 13 minutes to the first human contact, and 26% of firms never responded (JustCall summary). In home services, ServiceTitan’s 2026 survey, also a vendor claim, found only 50% of contractors respond to new leads within the first hour.

Work out your own number

You can estimate your exposure with four inputs: monthly inbound calls, your miss rate, the share of callers who are real prospects, and your average job value. Your phone system or carrier can usually give you the first two in a few minutes.

Here is the arithmetic using the Invoca averages. Take an HVAC company with 300 inbound calls a month and a 48% miss rate. That is about 144 unanswered calls. If 38% of them are qualified, that is about 55 real prospects. If 42% of those would have booked had someone answered, that is about 23 jobs a month.

This is arithmetic, not a result. It assumes callers who reach an AI behave like callers who reach a person, and nobody has measured that independently. Our research found no independent, quantified case study of an AI receptionist’s recovered bookings in home services, healthcare, law or home care. If someone quotes you a recovered-revenue figure as though it were proven, ask where it came from.

What an AI receptionist actually does

An AI receptionist answers your phone in your business’s name when your team can’t. In the systems we build, that usually means after hours, at lunch, and whenever the front desk is already on a call.

  • It opens every call by saying it’s an AI and that the call is recorded, in a line you approve.
  • It asks the questions your front desk asks, like what the problem is, the address, and how urgent it is.
  • It checks your calendar and books the job, or takes a message if you prefer.
  • It sends urgent calls straight to your on-call phone, and transfers to a person whenever a caller asks for one.
  • It connects to your phone number, your Google or Outlook calendar, and booking tools like ServiceTitan, Jobber or Calendly.

The handoff matters more than most owners expect. In a 2025 report from Five9, a company that sells contact-center software, 72% of consumers said they were open to AI if they could reach a human quickly (Call Centre Helper). The same company’s 2024 study found 75% prefer a real person (Business Wire). Both can be true. People prefer a person, and they would rather get a quick answer than a voicemail greeting.

What it does not do

It does not replace your front desk. It covers the calls nobody was free to answer. Your team still handles the conversations that need judgment, and the AI passes those over with what the caller already said.

It should not quote prices, promise coverage or give advice you haven’t approved. A business is responsible for what its AI tells customers. In 2024 a Canadian tribunal held Air Canada liable for wrong fare information its chatbot gave a customer (Moffatt v. Air Canada). We keep the AI to answers you’ve signed off on.

And it is not flawless on the phone. In an independent benchmark of 499 live calls, Codebridge found well-tuned systems took about 0.75 to 1.3 seconds to start replying. A caller pausing mid-sentence for 680 milliseconds tripped up every system tested, with 0 clean passes in more than 65 attempts. One fast, small model garbled a confirmation code 7 times out of 7.

That is why we test with real scenarios from your business before launch, and why the AI reads back phone numbers, dates and addresses before it books anything.

How to tell whether it’s working

Measure before you start. Pull a month of call records and note how many calls were missed, and at what times. After launch, compare calls answered after hours, bookings made, urgent calls transferred and messages taken.

We read the conversations it couldn’t finish every month, fix what it got wrong, and send you a short report. If the numbers don’t move, the report will show that too.

Next step

Tell us how many calls you get and when they go to voicemail. Start the chat and we’ll sketch what an AI receptionist would handle for your business. To see how it works alongside website chat and follow-ups, visit our front desk and sales page.