Insights / Phones, front desk and patient messages

Medical answering service vs AI receptionist: how to choose

Medical answering service vs AI receptionist for a practice: after-hours and urgent calls, costs, BAAs under HIPAA, TCPA rules and questions to ask.

When the front desk cannot answer every call, most practices end up looking at two options. One is a medical answering service: trained people, usually in a call center, who answer your line when you cannot. The other is an AI receptionist: software that answers with a computer-generated voice, talks with the caller and can often book or pass on a message. Many practices also combine them.

This guide compares them on the things that decide the choice for an independent practice: the calls you actually get, urgent and after-hours calls, cost, privacy contracts and the phone rules. It does not recommend a vendor.

This is general information, not legal advice. Rules cited were checked on October 27, 2026 and can change. Have your counsel review any vendor contract and how you handle urgent calls. We do not quote prices or answer rates, because they vary by vendor and we have no reliable public source for typical figures.

Start with your own calls

Before comparing services, find out what callers want. Pull a few weeks of call logs from your phone system, or have staff tally calls for a week by reason and time of day. Most practices find a handful of reasons cover most calls:

  • New patients asking whether you take their insurance and when they can be seen.
  • Existing patients booking, moving or cancelling an appointment.
  • Prescription refills, results and billing questions that need someone with access to the chart or the ledger.
  • Clinical concerns, from a simple question to something urgent.
  • Calls outside office hours and calls that hit voicemail during busy times.

The mix decides the answer. A practice whose missed calls are mostly booking requests at lunch has a different problem from one whose after-hours line gets clinical calls every night.

What a medical answering service does

A live answering service puts a person on the line. Typical work includes taking messages, sending them to the practice by an agreed method, reaching the on-call provider for urgent calls using a script you approve, and sometimes booking appointments in your system if you give them access.

Its strengths are human judgment on calls that do not fit a script, and comfort for callers who prefer a person. Its limits are the ones you can check: how long callers wait at busy times, how well agents know your practice, and whether messages are accurate and complete. Ask for sample messages and listen to a test call yourself.

What an AI receptionist does

An AI receptionist answers every call at once, at any hour, in the same way each time. Depending on the product, it can answer common questions from information you give it, book or move appointments in your scheduling system, take a structured message and transfer the caller to staff.

Its strengths are availability and consistency on routine calls. Its limits are calls it does not understand, callers who will not talk to software, and anything that needs clinical judgment. What matters is how quickly and cleanly it hands those calls to a person. Our guide to an AI receptionist for a medical office covers how to test the hand-off and booking in detail.

Side by side

  1. Routine booking and questions. Both can handle them. AI answers immediately at any volume; a live service depends on staffing.
  2. Urgent and clinical calls. Neither should give medical advice. A live service follows your escalation script with human judgment; AI must recognize the call and transfer or escalate reliably. Test both with real examples.
  3. Callers who want a person. A live service already is one. With AI, check that a caller can reach a person easily.
  4. Access to your systems. Both may need access to scheduling. Each connection is a privacy question; see below.
  5. Consistency and records. Ask each vendor what is recorded, transcribed and stored, for how long and where.

After-hours and urgent calls

This is the part that matters most, and the part to design yourself rather than leave to the vendor. Write down which situations go straight to the on-call provider, which can wait for a message the next morning, and what callers hear first. A common opening tells anyone with a medical emergency to hang up and call 911.

Then test it. Call the line after hours with a few scenarios, some routine and some that should escalate, and check what happened, how long it took and what the provider received. Repeat after any change to the script or the software. Whichever option you choose, the practice remains responsible for how urgent calls are handled.

How to compare costs

Pricing models differ, so compare them on your own call volume rather than on a headline figure. Ask each vendor, in writing:

  • How they bill: per minute, per call, a monthly plan with included usage, or a mix.
  • What counts as billable: hold time, transfers, short calls, after-hours calls and outbound calls.
  • Set-up, integration and cancellation fees, and the contract length.
  • What happens to the price if your volume grows or drops.

Then price a typical month using your call log from the first step. Include the staff time each option saves or adds, such as returning messages the next morning.

Privacy: business associate agreements

Under HIPAA, a business associate is a person or company that creates, receives, maintains or transmits protected health information on behalf of a covered entity (45 CFR 160.103). An answering service that takes patient messages, or an AI receptionist that hears, transcribes or stores what patients say, handles that kind of information.

HHS explains that only a mere conduit, such as the postal service, which transports information without accessing it except rarely, is not a business associate. A vendor that answers your patients’ calls and records their details is doing more than that. Where a vendor is a business associate, the Privacy Rule requires a written contract with the protections in 45 CFR 164.504(e), and HHS’s cloud computing guidance applies the same idea to companies that store data even if they do not read it.

So ask both kinds of vendor whether they will sign a business associate agreement, which subcontractors handle call audio and transcripts, and how data is protected and deleted. A signed agreement is a requirement, not a guarantee; how you set up and use the service still matters. Our guide to practice email and fax under HIPAA goes through the same vendor questions for other tools.

Outbound calls and the TCPA

Answering incoming calls is one thing. Many AI tools can also make calls, such as reminders or recall calls. On February 8, 2024, the FCC issued a declaratory ruling (FCC 24-17) confirming that calls using AI-generated voices are calls with an artificial voice under the Telephone Consumer Protection Act. The FCC’s consent rules for artificial or prerecorded voice calls, in 47 CFR 64.1200, therefore apply to them.

Those rules include a specific provision for certain healthcare calls, with conditions. If you plan to use AI for outbound calls, have counsel confirm which of your calls qualify and what consent you need before you turn it on. Our patient recall guide covers reminder and recall messages.

Questions to ask before you sign

  1. Can we hear real calls? A test line or recorded examples with our own scenarios.
  2. How do urgent calls escalate, and how fast does the on-call provider get them?
  3. Will you sign a business associate agreement? Which subcontractors handle our patients’ information?
  4. What is recorded or transcribed, where is it stored, and for how long?
  5. How does a caller reach a person, and what happens if no one picks up?
  6. What does a month cost at our real call volume, and how do we leave?

If you want help working through this for your practice, our AI front desk service page explains how we approach it, and the new patient phone script helps whichever option answers the phone.

Frequently asked questions

Is an AI receptionist cheaper than an answering service?

It depends on the vendors’ pricing and your call volume. Ask each one how they bill and what counts as billable, then price a typical month using your own call log, including staff time saved or added.

Can an AI receptionist handle after-hours medical calls?

It can answer and route them, but it should not give medical advice. You decide which calls go to the on-call provider, and you should test the escalation yourself with realistic scenarios.

Does a medical answering service need a business associate agreement?

A service that takes and records patient information on your behalf generally fits the HIPAA definition of a business associate in 45 CFR 160.103, which requires a written agreement under 45 CFR 164.504(e). Have counsel review the contract.

Does an AI receptionist need a business associate agreement?

If it receives, transcribes or stores patient information for your practice, the same definition applies. Ask the vendor to sign one and to name the subcontractors that handle call data.

Can AI make reminder calls to patients?

The FCC ruled in February 2024 (FCC 24-17) that AI-generated voices are artificial voices under the TCPA, so the consent rules in 47 CFR 64.1200 apply. Some healthcare calls are covered by a specific provision with conditions; confirm with counsel.

Can we use both?

Yes. Some practices use AI for routine calls and a live service or on-call staff for urgent ones. Write down which calls go where and test the hand-off between them.

If either option records or transcribes calls, see our call recording consent guide for greetings, state consent rules and retention.

Sources

More in Phones, front desk and patient messages

Want a second pair of eyes on your patient path?

Book a practice review. We look at how patients find you, what they see on your site, what happens when they call, and what you can measure today. You get a plain list of what we would fix first.