September 4, 2026
What an AI receptionist actually costs
Paige Barrett
September 4, 2026
Paige Barrett
Vendors price this four different ways, and the model matters more than the headline number, because two quotes that look similar can differ substantially once a med spa's real call volume runs through them. This guide covers the four pricing models and which kind of aesthetic practice each one suits, the costs that do not appear on the quote, how to compare against the alternatives a clinic is actually choosing between, and how to work out what phone coverage is worth against your own treatment values before you take a single sales call. We do not publish our own pricing, so what follows is the framework rather than a figure.
Per minute. You pay for connected call time. Common with services that grew out of traditional answering, and it suits practices with low volume and short calls. It becomes unpredictable when calls run long, which is exactly what happens when a system is good enough to handle a full booking conversation. Worth noting the incentive: a per-minute model does not reward brevity.
Per call. A fixed amount per answered call, regardless of length. Easier to forecast than per minute and it stops long conversations from being penalized. Check what counts as a call. Wrong numbers, hang-ups, robocalls, and the same patient ringing back twice about a consultation may all be billable, and in a clinic running heavy paid social with the spam volume that attracts, that gap is significant.
Flat monthly. A fixed fee, usually with a usage allowance and an overage rate above it. The most predictable and the easiest to budget, which is why most practice software is sold this way. The thing to check is the allowance and what happens above it, because a flat fee with a low ceiling behaves like a per-call model once you cross it.
Per seat or per location. Priced by provider, user, or site. Common where the tool is part of a wider platform. Suits multi-location groups and practices running several injectors. It is poor value for a solo injector paying for capacity they will not use.
The right question for any quote is what your clinic's last three months of call volume would have cost under that model, including the after-hours consultation inquiries you are currently missing, since those will now be answered and will raise billable volume. Which model is cheapest in principle tells you nothing.
Six things move a quote, and knowing them lets you read one properly.
Coverage hours. After-hours only costs less than around the clock, which costs less than full coverage including overflow while your injectors are in treatment rooms. Decide what you actually need before you ask, because it is the largest single lever.
Call volume. Under any model, more calls cost more. Use your real inbound volume from your phone system, not an estimate, and remember to include the calls currently going unanswered.
Capability. Answering and taking a message is the cheapest tier. Answering questions from your own content costs more. Booking into your live schedule costs more again, because it requires a real integration. This is the biggest source of quotes that look wildly different for what appears to be the same product.
Integration depth. Reading and writing to your practice management system, EMR, or treatment calendar is real engineering. A tool that only sends the front desk an email costs less because it does less.
Number of locations and providers. Multi-site routing, and rules for which injector or surgeon can be booked for which treatment, add complexity and usually cost.
Setup and training. Configuring a system on your treatment menu, policies, hours, and booking rules is either included, charged once, or charged as a project. This is the line most often missing from a comparison.
Four of these trip practices up, and all four are askable in a first call.
Implementation. Whether setup is included, one-time, or billed by the hour. Ask how many hours of your team's time it takes as well, because that is a real cost even when the vendor charges nothing.
Integration fees. Some platforms charge for the connection to your practice management system, and some charge on both sides. Confirm your PMS supports the integration before you sign anything, since a system that cannot write to your schedule cannot book.
Overage. The rate above the allowance, and how it is calculated. Ask for the overage rate in writing and model a busy month, not an average one.
Contract length and exit. Whether it is monthly or annual, what happens if you leave mid-term, and how you get your call records and transcripts out. A tool holding a year of patient conversations should not make them hard to retrieve.
Two more worth asking about: what ongoing changes cost when your services or hours change, and whether support is included or a paid tier. A system your team cannot update without a vendor ticket will drift out of date, and the cost of that shows up as wrong answers to patients rather than as a line on an invoice.
The comparison decides whether the number is high or low, and most practices compare against the wrong thing.
Against a hire. Count the full cost of a front desk person: salary, payroll taxes, benefits, recruiting, training time before they are productive, and the coverage that disappears when they are sick or on vacation. A hire also covers a shift, so the phones are still unanswered at 9 PM and on Sunday. A hire and automated coverage solve overlapping but different problems, and many practices end up with both.
Against your current answering service. A closer comparison, and the question is what you get for the difference. If the current service takes messages and the alternative books consultations, you are comparing a Monday callback queue against filled slots, which is a difference in outcome rather than in price.
Against voicemail. This is the real comparison for after-hours, and it makes the math simple. Voicemail costs you the bookings that did not happen.
Do this before you take a sales call. It takes fifteen minutes and it changes the conversation.
The number that comes out is usually decisive one way or the other, and it belongs to your practice rather than to a case study. A practice missing a handful of calls a month reaches a different conclusion than one missing several a day, and both conclusions are correct.
One caution on the arithmetic. Do not assume every answered call converts. A reasonable model recovers a share of missed calls, not all of them, and some callers would have called back anyway. Run it at a conservative recovery rate first. If it still works at conservative, it works.
Often more than for a large one, which is the opposite of what most people assume.
A large practice with a full front desk team already answers most of its calls during business hours. The gap is after hours and overflow. A solo injector or a two-person practice has a gap that covers most of the working day, because the person who would answer the phone is treating a patient, and every call missed during clinic hours is a call that reached a competitor with a receptionist.
The counterweight is volume. A very small practice may miss few enough calls that no paid option pays for itself, and the honest answer in that case is to stay with voicemail and revisit when volume grows. The arithmetic above tells you which situation you are in. Run it before assuming either.
Recura is the AI receptionist that is part of the PatientNow family, and we do not publish pricing, because what a practice needs varies enough that a headline number would mislead more often than it helped. Coverage hours, call volume, whether you want booking or message-taking, and how many locations and providers you run all move it.
What we will do on a first call is size it against your actual numbers rather than a package. Bring your call log and your average treatment value, and the conversation is about whether this pays for itself in your practice, which is the only version of the question that matters.
Both models exist, along with per-minute and per-seat. Flat monthly with a usage allowance is the most common for software-based systems and the easiest to budget. Per-minute and per-call are more common with services that grew out of traditional answering. Ask which model a vendor uses before comparing any headline figure, because the models are not comparable on price alone.
It varies. Some include configuration, some charge a one-time implementation fee, and some bill it as a project, particularly where an integration with your practice management system is involved. Ask specifically, and ask how much of your team's time is required, since that is a cost even when the vendor charges nothing for it.
Under a flat monthly plan there is normally an overage rate above the included allowance. Get that rate in writing and model a busy month rather than an average one. This is the most common source of a bill that is larger than expected, especially in the first months, when previously missed calls start being answered and volume rises.
Usually yes, and a sensible way to begin. After-hours coverage is the clearest gap, the easiest to measure, and the lowest risk because those calls are currently going to voicemail. Once you can see how it handles them, widening to overflow during business hours is a smaller decision.
If your gap is daytime volume that needs human judgment, a hire is often the better answer and better service. If your gap is coverage at hours nobody is rostered for, a hire does not close it at any salary. Most practices past a certain size need both, and the useful exercise is separating which hours need a person from which hours need coverage.
Bring your call log and your average treatment value to a first call, and we will size it against your numbers rather than a package.
Recura answers calls, texts back leads, and books consults, so your calendar fills even overnight.
See a live call flow, tailored to your practice
Recura insights and updates, shared when there’s something worth knowing.
Proudly part of the
family
