An owner staffing a phone line has three real options: hire someone to answer it, pay an answering service to pick up on the business's behalf, or put an AI system on it. Almost nobody runs the numbers past whichever one they tried first. Here's that math, with every input sourced and the arithmetic shown, including our own prices.
What a receptionist actually costs
The Bureau of Labor Statistics puts the median wage for a receptionist at $18.27 an hour, $38,010 a year, as of May 2025. That's the wage. It isn't what the job costs the business.
BLS's separate compensation data (the Employer Costs for Employee Compensation release, March 2025) shows benefits running 29.7% of total private-industry compensation, with wages making up the other 70.3%. Take the $38,010 wage as that 70.3% share and solve for the whole: $38,010 ÷ 0.703 ≈ $54,068 a year, fully loaded. That figure is our arithmetic, not a published BLS line item. It's worth showing because most cost comparisons skip this step and put a vendor's price tag next to a bare salary nobody actually pays.
That $54,068 buys one person on one shift. Roughly 40 hours a week, business hours, minus lunch, minus sick days, minus whatever they're doing when two lines ring at once. Nights and weekends aren't in that number. Covering them means overtime, a second hire, or both.
What a live answering service costs
Answering services bill by the minute, typically $0.65 to $2.50 depending on the vendor and plan. Assume a 3-minute call, a reasonable middle estimate for a triage conversation, and that works out to roughly $1.95 to $7.50 per call. Monthly plans from the larger services package the same economics, running a few hundred dollars a month for a set block of minutes.
The structural difference from a hire: no fixed cost and no ceiling. A slow month costs almost nothing. A busy one costs proportionally more, and the per-call rate never drops as volume grows.
Where hiring and the answering service cross
Set the fixed $54,068 hire against the answering service's per-call cost and solve for the volume where they match.
At $1.95 a call: $54,068 ÷ $1.95 ≈ 27,727 calls a year, about 76 a day.
At $7.50 a call: $54,068 ÷ $7.50 ≈ 7,209 calls a year, about 20 a day.
Below 20 to 76 calls a day, depending on which end of the rate card applies, the answering service costs less than a salary. Above it, hiring wins on paper. The catch is in "on paper": at the top of that range, one person would be answering a call every six minutes of an eight-hour shift with time for nothing else. The crossover on paper and the crossover in a shop that still answers its phone are two different numbers.
What the $50 AI plan leaves out
Self-serve AI answering apps publish plans from under $50 to a few hundred dollars a month. Those prices are real, and for some businesses they're enough. But the sticker price covers the call itself. Before comparing it to anything, check what happens after the call ends:
- Who sets it up, writes its scripts, and keeps them current when your services or prices change?
- Does the caller's information land in your CRM, or in an inbox someone has to copy from?
- Does it book straight onto a real calendar, or take a message?
- Who follows up on the quote that went quiet, or the appointment that no-showed?
- Who do you call when it starts giving wrong answers?
If the answer to most of those is "you," the app is cheap because you're the rest of the system. That's a fine trade for a one-truck operation with time on evenings. It's an expensive one for a shop where the owner's hour is worth more than the plan.
What we charge, and what's in it
Level 99's phone coverage starts at our LVL 10 plan: $699 a month plus a $797 setup fee. (Our $349 entry plan covers text and web leads only, no phone agent, so it isn't the fair comparison here.) That price includes the voice agent plus the system around it: CRM configuration, lead intake and routing, appointment booking and confirmation, automated follow-up and no-show recovery, and pipeline rules for handing leads to your team. Our LVL 50 plan, $1,199 a month plus $1,497 setup, adds outbound calling and a third agent. Every tier is on the pricing page.
Run it through the same math. Year one on LVL 10 is $699 × 12 + $797 = $9,185, about 17% of the $54,068 fully loaded receptionist, with nights and weekends covered. Against the answering service, $699 a month equals the cost of roughly 93 calls at $7.50 or 358 calls at $1.95. Above that range, a flat monthly price beats per-call billing, provided the monthly price actually covers your volume. Ask any vendor what happens above their plan's limits, us included.
The number cost-per-call doesn't show
An Aire Serv franchise ran a live answering service against Avoca's AI voice agent over the same stretch of calls. This is vendor-published data from one named franchise, not an independent study, and it should be read that way. The live service booked 5 jobs. The AI booked 43.
The gap there has nothing to do with price. It comes from answer rate and follow-through, the same failure mode covered in what actually happens to unanswered calls on a trades line and in what changed for FrontPorch when voicemail got replaced. One data point from one franchise doesn't prove the gap holds everywhere. It does mean a comparison that stops at the price sheet measures the wrong thing when the options don't book jobs at the same rate.
No independent research exists on how fast any of these options pays for itself, so here's the arithmetic instead of a borrowed number. Say a shop's average job is worth $400. On LVL 10, two additional booked jobs a month cover the $699, and the $797 setup is two more jobs, once. Whether that happens depends on how many calls the shop is missing today, and that number should come from its own phone line, not from a vendor's pricing page, ours included.
FAQ
Can one receptionist really handle 20 to 76 calls a day?
At the low end, plausibly. At the high end, no. That's a call every six minutes across an eight-hour shift, with no time for routing calls, taking messages, or handling walk-ins. The crossover math shows where hiring is cheaper on paper, not where one person can staff it.
Why would an answering service ever cost more than hiring someone?
Per-minute billing has no ceiling. A salary is fixed; a per-minute vendor keeps charging as volume rises, so at high enough call counts the math flips, even though the service felt cheap at low volume.
Isn't a $50 AI app good enough?
For some businesses, yes. If you have the time to set it up, keep its answers current, and do the follow-up yourself, it can be the right call. The $50 buys the answering. The CRM updates, the booking, and the follow-up on every lead that doesn't close on the first call are either built in by someone or done by you.
Is the Aire Serv 5-versus-43 result typical?
There's no way to know from one published test. It's a real number from a named franchise, worth factoring in, not a promise that any given shop will see that margin. Track your own booked calls before and after any change, whichever option you pick.