What an Answering Service Really Costs (and Where the Money Leaks)

A woman calls a dental office I worked with at 6:40 on a Tuesday evening. She has a cracked molar and a dental plan she does not understand. The office closed at 5. Her call rolls to the answering service, where an operator three states away reads from a card that says "we are closed, please call back during business hours, or I can take a message." She hangs up and calls the practice down the road that picks up. That call cost the office nothing in answering service fees. It cost them a patient worth about four thousand dollars over the next two years.

That is the thing nobody puts on the invoice. So let me walk through what these services actually charge, line by line, because the sticker price is rarely the real price.

How traditional answering services bill you

Most live answering services price one of three ways, and a lot of them blend all three to make comparison hard.

Per minute. This is the common one. You pay for time the operator spends on your account. Rates I have seen run roughly $0.80 to $1.50 a minute. Sounds cheap until you remember a real call is not 60 seconds. By the time the operator greets the caller, pulls up your script, asks the questions, repeats the spelling of an email twice, and wraps up, you are looking at three to five minutes. Call it $3 to $7 per call, before anything else hits.

Per call. Some services quote a flat per-call rate, often $1.50 to $4. Cleaner to read, but watch the definition of "call." A wrong number that the operator picks up and dispatches as a message can count. A caller who hangs up after the greeting can count. You are paying for the connection, not the outcome.

Monthly minimum or retainer. Almost all of them carry a base. I have seen $50 a month on the low end for a tiny plan and $200 to $500 for anything with real volume. The minimum is the floor you pay even in a slow month, and slow months are exactly when you can least afford a fixed bill.

The fees that do not make the headline

The quoted rate is the lure. Here is what tends to show up once you are a customer:

• Setup or account fees, often $50 to $150 to build your script.
• Overage rates that jump once you blow past your bundled minutes, sometimes double the base.
• Holiday and after-hours surcharges. The exact hours you most need coverage cost more.
• Per-message or per-patch fees for texting or transferring a call to you.
• Billing increments that round up. A 65-second call billed in full-minute blocks is two minutes. Over a month that rounding is real money.

I once helped a home services shop audit their answering service bill. The advertised rate was $1.19 a minute. Their effective rate, after minimums, rounding, and patch fees, worked out closer to $2.40. They were paying nearly double the number on the brochure and had no idea.

What you are actually buying

Cost only makes sense next to value, and live services have a quality ceiling you should price in.

The operator is not from your business. They read a card. They cannot answer "do you take Delta Dental PPO" or "what does a water heater swap run." They take a message and promise someone will call back. For simple message-taking after hours, fine. For anything where the caller wants an answer or wants to book, the operator becomes an expensive voicemail with a pulse. Hold times during call spikes are common because one operator juggles many accounts. And almost none of them touch your other channels. The website chat, the texts, the email that comes in at 9 p.m. all sit untouched.

How usage-based AI pricing compares

This is where the math changes shape. With LastWorker there is no monthly fee and no minimum. You load a prepaid balance and pay only for conversations it actually handles. Voice is billed per second at $0.05 a minute, chat and SMS per message, email per resolved ticket. If you want a dedicated number it is $1 a month. Auto-reload tops up the balance when it runs low so coverage never drops.

Per second matters more than it sounds. A 65-second call is 65 seconds, not two minutes. No rounding tax. No after-hours surcharge, because 2 a.m. costs the same as 2 p.m. No setup fee, since you configure it in about a fifteen-minute conversation where it learns your services, pricing, hours, and policies. No code.

Here is the same evening call, priced both ways.

The AI does not take a message and hope. It answers the question, books or reschedules the appointment, captures the lead, and transfers or escalates to a human when something genuinely needs one. It does that in 97 languages with replies under a second that sound human. The cracked-molar caller gets an appointment instead of a dial tone.

Run your own numbers

Pull your last three answering service invoices. Find the all-in monthly total, then divide by the number of calls that produced an actual outcome you cared about, a booking, a qualified lead, a real message. Not the wrong numbers. Not the hang-ups. That number is your true cost per useful call. Most shops I have worked with are stunned by it, usually somewhere between $4 and $9.

Then count the calls, chats, and texts that never got answered at all, because those are free on the invoice and brutally expensive in the bank. Across the businesses I have run, missing roughly a quarter of inbound contact is normal, and after-hours is where most of it hides.

If you want to see the full per-channel breakdown, the pricing page lays out voice, chat, SMS, and email rates plainly. And if you are weighing this against a specific human service, the comparison pages put them side by side.

A live answering service made sense when the only alternative was a machine that beeped. That is no longer the alternative. The honest comparison today is not cheap operator versus expensive AI. It is a per-minute message-taker with a monthly floor versus a per-conversation system that actually closes the loop and costs pennies a call. I have signed the checks for both. The cheaper line item is not always the one with the smaller advertised rate, and the cracked molar at 6:40 is the call that proves it.