What does an AI receptionist cost?
Five different pricing shapes, none of them comparable as quoted. Here’s how to convert them — and why the number that matters is cost per successfully handled call.
The short answer, and the better question
AI receptionist pricing is quoted in at least five different units, which is why comparing two quotes feels like comparing a mortgage to a taxi fare. Before you can compare prices at all, you have to convert them to the same unit — and the unit that actually matters isn’t any of the five.
It’s cost per successfully handled call. A cheap agent that fumbles a third of its bookings is not cheap; it is a discount on the calls it handled and a full-price disaster on the ones it lost. That’s the number this page is about, and it’s the only pricing question we can help with that nobody else can — because it needs a measurement of quality, and quality is what we do.
The five pricing shapes you’ll be quoted
- Per minute. The most common shape, inherited from telephony. Watch the rounding — per-second, per-six-seconds and per-minute billing produce meaningfully different invoices on short calls, and most front-desk calls are short. Also ask what happens to hold time, transfer time and the seconds after the caller hangs up.
- Per call or per conversation. Simpler to forecast, and it stops rewarding a slow agent. Ask what counts: does a 4-second wrong number bill? Does a call the agent transfers to a human bill once or twice? Does a caller who rings back after a failure bill again?
- Per seat or per location. Priced like software rather than telephony. Fine for a single practice; ask how a second location, a second phone number or an after-hours overflow line is counted before you grow into a surprise.
- Flat monthly. Usually flat up to an included volume, then an overage rate that is the actual price you should be comparing. Ask for the overage rate first — it is the number that governs a busy month.
- Setup and onboarding fees. One-time build, integration or training charges. Legitimate work, but they change the arithmetic of a short pilot enormously and are the easiest line to forget in a comparison.
Underneath any of them, ask who pays for the phone numbers and the carrier minutes, whether an annual commitment is required for the quoted rate, and what the notice period is. A quote that only exists at twelve months is a different product from one you can leave in thirty days.
Convert every quote to one number
Do this on the back of an envelope before your next vendor call:
- Fix a month of your real call volume. Not the vendor’s example — yours. Pull the count off your phone bill.
- Price that month under each quote, including setup amortised over the term you’d actually commit to, plus the overage rate for a busy month rather than an average one.
- Divide by the calls that would actually go right. This is the step everyone skips, and it is the whole game.
An illustration, with made-up round numbers purely to show the arithmetic — these are not measurements of any vendor: two agents, a thousand calls a month. One costs a third less per call than the other but completes eight of ten tasks where the other completes nine and a half. Per call, the cheap one wins easily. Per successfully handled call, it costs more — and that comparison still hasn’t counted the two hundred callers who didn’t get what they rang for.
What a failed call actually costs
The reason the arithmetic tilts so hard is that the two sides of it aren’t the same size. The saving on a call is measured in cents. The cost of a lost one is measured in whatever a customer is worth to you: a new patient who books elsewhere, an emergency job that went to the next number on the list, a reservation that became a bad review.
Which is why we would tell almost any buyer to grade first, then compare price among the vendors that pass — free report cards to narrow the field, then a run against your own situations to choose between the two that survive. Price differences between AI receptionist vendors are real but bounded. Quality differences are not — in our testing, the gap between the best and worst handling of the same ordinary call is far larger than the gap between their prices.
What to ask the vendor about price
- “What exactly bills?” Wrong numbers, hang-ups in the first five seconds, transfers to a human, voicemails, failed calls the caller has to repeat.
- “What’s the overage rate?” Then price your busiest month at it, not your average one.
- “What’s the all-in first-year number?” Setup, integration, numbers, minutes, support tier, the lot.
- “Of your last thousand calls, how many ended with the caller hanging up unserved?” A vendor who can’t answer isn’t measuring — and a vendor who isn’t measuring can’t tell you what you’re actually buying. That question, and nine others, are in the ten-question checklist.
Common questions
How much does an AI receptionist cost?
Vendors quote in at least five different units — per minute, per call, per seat, flat monthly with overage, and one-time setup — which is why two quotes rarely compare directly. Convert each one to the cost of a month of YOUR real call volume, then divide by the calls that would actually go right.
Is an AI receptionist cheaper than a human answering service?
Usually per call, yes, since human services are typically priced per minute with a monthly minimum. But the comparison only means something once you know how many calls each one actually completes — which is what our grades measure.
Does a higher price mean a better AI receptionist?
It predicts quality poorly. We have graded vendors across the price range and found the gap in call handling far wider than the gap in price. Grade first, then compare price among the vendors that pass.
What’s a hidden cost people miss?
The overage rate on a flat plan, setup fees amortised over a short pilot, per-location charges when a second office arrives, and — largest of all — the revenue on calls the agent didn’t complete.
Do you publish vendor prices?
No. Prices change constantly and are frequently negotiated, so a price table here would be wrong within weeks. We publish the thing that doesn’t come off a rate card: how well each vendor actually handles the calls.
The independent guide to AI receptionists. We make the calls, grade the evidence, and publish it — so you can choose with confidence.