Answering service vs hiring a receptionist: which makes sense, and when?

There's a point where the phone stops being background noise and becomes the thing holding the business back. You're returning voicemails at nine at night. Two calls landed while you were on a job and you got neither. So you price the two obvious fixes: put somebody at a desk, or hand the line to an answering service.

Most owners run that as a cost comparison: a salary on one side, a monthly invoice on the other, pick the cheaper. That's the wrong axis, and it's how shops buy the option that doesn't fit them.

Here's the short version. Hire a receptionist when your phone is busy enough through the business day to keep one person occupied, and when whoever answers has to know your pricing, your service area and your regulars. Use an answering service when the calls that pay land at night, on weekends, and in bursts you can't staff for. A receptionist buys depth across about 40 hours a week. An answering service buys coverage across all 168. Neither buys both, and that trade is the real decision.

Why is this a coverage question and not a cost question?

The arithmetic isn't news. A week holds 168 hours and a full-time seat covers 40, fewer once you count lunch, a dentist appointment and the week in July they're on vacation. Call it 35 in a good week.

What gets skipped is the other 130. That's no knock on the person you'd hire. One person holds one conversation, on one shift. It's arithmetic, not attitude.

To be fair to the hire, nobody needs volume coverage at three on a Tuesday morning. The 168 isn't a target, it's the denominator. What matters is which of those hours your revenue calls land in, and the four questions below are how to find out.

An answering service inverts the problem. Somebody picks up at two on a Sunday morning, which a desk will never do. What they start with is a script, not your business. Your calendar, your service-area edges and your pricing get in only as far as you build them into it and pay for the tier that carries it, and none of it tells them the caller is your customer of nine years. Fuller treatment in AI receptionist vs answering service.

What happens to the same call under each?

Take one call and run it both ways. It's the Friday before a holiday weekend, 4:40, and the first heat wave of the year is two days old. A homeowner's system quit overnight, the condenser is fourteen years old, and both techs are out on other jobs.

With a receptionist, at 4:40, this goes well. They pick up on the second ring, know Saturday morning is workable, and know to ask the age of the unit, so a fourteen-year-old condenser quitting in a heat wave reaches you flagged as a replacement conversation and not one more repair slot. The job is on the calendar and the details are with your lead tech before the homeowner hangs up. That's the best outcome anywhere in this post, and worth saying plainly.

Now move the same call to 5:40. The desk is dark and it rings out to voicemail, where Invoca's own platform data puts the share of callers who leave a message at fewer than 3%. You don't get a thin lead. You get nothing, and never learn the call happened. More in do customers actually leave voicemails anymore.

With an answering service, 5:40 gets answered. What happens next is whatever you set up. On a good account with a real emergency protocol, no cool in a heat wave pages your on-call tech and your phone buzzes within a minute. On the entry-level message-taking version, somebody takes a name, a number and "no AC," and you read it at eight.

That second version is where the money leaks. Researchers who looked at 1.25 million sales leads, reported by Oldroyd and colleagues in Harvard Business Review in 2011, found firms that made contact within an hour were about seven times as likely to qualify the lead as firms that waited even an hour longer.

Those were web-form leads at B2C and B2B companies, not a homeowner in a hot house, so take it as direction, not as your number. Nobody sweats through a holiday weekend wondering whether you'll ring back.

Which four questions actually decide it?

Four things settle it, and your own call history answers all four.

• When do your money calls actually land? Pull ninety days and split it at eight to five, Monday to Friday. If most of your revenue starts inside that window, a hire is genuinely on the table. If it starts outside it, you can't staff your way to the fix.
• What happens when two calls arrive together? One person holds one conversation. The second caller can be parked, and parked isn't answered: a homeowner on hold in a heat wave is a homeowner dialing the next name. A service's queue does the same when the whole county calls at once.
• How much does the answerer need to know? Insurance work, service-area edges, quoting that turns on what the roof or the panel looks like: that's depth, and depth lives in-house. Straight triage and booking is work an outside service does well.
• Does the call end on your calendar, or as a message you still have to work? This is the one that gets skipped. A message is a lead you haven't caught yet, and it ages by the hour.

What does each one really cost you?

On the hire side, start with the wage. BLS OEWS 2025 figures put median receptionist pay at $18.27 an hour, about $38,010 a year. That's a national median, so in a metro market plan on being above it.

The wage isn't the cost, though. Add payroll taxes, workers comp, a desk and a phone, and the weeks it takes anybody to learn your trade well enough to tell a maintenance call from an emergency. Use your own rates, not a rule of thumb. Then add the coverage gap: one person covers one shift, so every hour outside it, holidays and vacation weeks included, rings into whatever you had before.

On the service side there's no wage, and the shape of the bill is what matters. They bill per minute, per call, or on a flat monthly tier with an overage above it. On a per-minute plan the invoice moves with your volume, so the month of the heat wave, the freeze or the hail storm brings the biggest bill. That's the month you can least forecast and most need covered.

You'll notice no monthly figure here, for the service side or for us. Ranges quoted in this category mean little without the plan shape behind them.

Neither option is expensive in the abstract. What's expensive is paying for coverage you don't get.

What's the honest case for putting a person at the desk?

A good receptionist isn't a phone answerer, and that's what the salary-versus-invoice comparison misses.

They're a front desk. They handle the walk-in, chase the parts order, rebuild the schedule when a tech calls in sick at six in the morning, and remember that Mrs. Alvarez's furnace goes out of warranty in November. Give it a year and a good one knows what closes in your market, because they hear every call.

One number should temper all of this. Invoca's 2026 report found that on 55% of the home services calls it scored, nobody asked the caller to buy or book the job. Those are calls a person picked up. Picking up and asking for the work are two different jobs, and the second doesn't come automatically with the first.

What a desk of your own gives you is somebody you can hold to that, who hears how every quote ended, and who your repeat customers come to know.

The honest case for the answering service is shorter and just as real: a trained human on your line at two in the morning beats voicemail every single time.

Where do both options break at the same time?

At the peak. Which is, inconveniently, the hour that pays.

That same Invoca report breaks answer rates out by sub-industry and puts the share of inbound calls to HVAC companies that reach a person at 34%, one of the lower rates among the nine it measures. That covers all inbound calls, misdials and hangups included, so it isn't a count of lost jobs. Read it as evidence the HVAC phone is hard to keep up with.

The reason is the shape of the demand. The first 95-degree Friday isn't a busier day than Thursday was, it's a different kind of day, where a season's worth of calls turns up inside a few hours. Roofing gets the same after hail, plumbing after a hard freeze. More in what happens to your overflow calls when every line rings at once.

A desk can hold one conversation in that hour. A service holds a queue, which is better, and a queue is still a wait, with the caller scrolling to the next name on the map while they sit in it. A big staffed service has real depth a single desk doesn't. Both still work against the same hour, and it's the hour that mattered most.

So which one should you pick?

Three rules, and most shops land cleanly on one of them.

• Calls mostly inside business hours, with real front-desk work besides the phone? Hire, and put something separate on nights and weekends.
• Calls mostly after hours, or arriving in bursts you can't schedule around? Don't hire first. Cover the hours.
• Already got somebody at the desk? Don't replace them. Cover the gaps they were never staffed to reach. More in how HVAC companies handle after-hours calls.

And whichever way you go, a call that ends as a message isn't a fix. It's the same problem, three hours later.

What if you didn't have to choose?

This is hard because both answers are partial. One gives depth without the hours, the other hours without the depth. You're picking which half of the problem to keep.

Willison is an AI receptionist built for the trades, and it's fair to say up front which half it takes. It takes the hours. Every call on your line gets answered, around the clock, in seconds. Ten calls landing in the same hour don't form a queue: they're answered at once, so the last caller waits no longer than the first.

It doesn't stop at the pickup. It asks the qualifying questions, puts the job on your calendar rather than in a message, and chases the quote that's gone quiet so a maybe doesn't become a never.

Now the half it doesn't take, since this post asked that of the other two. It won't remember Mrs. Alvarez, chase parts, or handle the walk-in. It doesn't dispatch a crew, send a truck, or promise an ETA. And it can't help a calendar that's already full. If you've got a front desk, this isn't a replacement. It's coverage on the hours that desk was never staffed to reach.

It's also real setup, the same as the booking tier you'd buy from a service: your hours, your service area and your pricing built in, your scheduler connected, before the first live call. The difference is who carries that. Willison manages it, the founder reviews it before the line goes live, and keeping it right never lands back on you.

Don't take the sound of it on faith. Press play on a real call on willisonhq.com and hear how it handles the call from first ring to booked.

Frequently asked questions

That depends entirely on what you buy, and it's the first thing to pin down in writing. Don't assume booking is included: ask whether the quote covers live booking into your scheduling system or only message taking, and what it takes to connect your calendar. Then ask exactly what lands on your phone after a call. A booked appointment and a message you still have to work at nine at night are very different products.

There's no clean threshold, because the right one depends on what else the seat covers. Run it backwards. Pull ninety days of call history, count only what landed between eight and five on a weekday, then ask whether that plus the rest of a front desk day (scheduling, parts, walk-ins, chasing quotes) fills the hours you'd be paying for. If the phone alone has to carry the case, it usually can't.

Almost certainly, and not as a replacement. Look at where the calls land that your desk was never staffed to reach: after five, weekends, the lunch hour, and the stretch when three calls arrive together and two ring out. Point coverage at those gaps only. The person at the desk keeps doing what a person is better at: everything needing judgment and a memory of your customers.