The calls a practice actually loses
A dental practice loses new patients in three predictable windows: at lunch when the front desk is covering a chair, after 5pm when people are finally free to make personal calls, and on Monday morning when weekend pain forces the issue. In each of those windows a new-patient caller will simply try the next practice on the results page.
Illustrative, using a first-year new-patient value of $600–$1,400. The exact figure matters less than the unit: this is measured in patients, not in minutes.
What a dental answering service costs in 2026
Human services bill per minute or per call, so the price rises with your busiest month rather than your best one. AI services bill a flat monthly fee. Judge both against the value of one recovered new patient, not against each other.
| Option | Typical monthly | Billed on | Books your calendar |
|---|---|---|---|
| Voicemail | $0 | — | No |
| Human answering service | $300–$1,200 | Per minute or call | Takes a message |
| Part-time front desk cover | $1,400–$2,600 | Hourly | Yes |
| AI answering service | Flat fee | Monthly | Yes |
A service costing $600 that recovers one $900 patient has paid for itself with four calls to spare. That is the only comparison that decides this.
What it should ask a new-patient caller
A new-patient call should be qualified in five questions. Longer and the caller disengages; shorter and the front desk has to call back anyway, which defeats the point.
- Is this an emergency, and what is the pain level? Decides whether the call is booked today or scheduled.
- New patient or existing? Different calendars, different appointment lengths.
- Which insurance carrier? The question most likely to end the call if answered wrong.
- What treatment do you think you need? In their words, not clinical terms.
- Best number to reach you? Confirmed back to them before the call ends.
If a vendor cannot show you the exact wording it will use for those five, they are selling you a voicemail box with better manners.
The recall list is where the money is
Every practice has patients who have not been in for twelve months or more. They are already in your system, already accepted your fees, and already know where to park. They are the cheapest appointments a practice can book — and almost nobody works the list, because working it means an afternoon on the phone being told no.
Answering protects the patients you were about to lose. Reactivation goes and gets the ones you already gave up on.
A 300-patient recall list worked properly converts at somewhere between 5% and 12%. At a $250 hygiene appointment that is $3,750 to $9,000 of chair time from a list you already own. This is usually the part that pays for the first month, and it is the part no phone system does.
Before you sign: six checks
- Ask for a signed BAA. Recordings and transcripts contain patient health information. A vendor who cannot produce a Business Associate Agreement is not a candidate.
- Ask where recordings live, and for how long. Storage location, retention period, and who on their side can open them.
- Ask whose phone number it is. If the number belongs to the vendor, leaving them costs you the number your patients have saved.
- Ask what happens on a transfer. Every service fails some calls. What matters is whether it hands over cleanly or hangs up.
- Get the escalation rules in writing. Which calls always reach a human: emergencies, complaints, anything above a price threshold.
- Ask to hear a real recording. Not a demo reel — a real call, with a real interruption in it.
When it is not worth it
An answering service is a bad purchase for a practice that already answers nearly every call. If your front desk is staffed through lunch, you do not run emergency hours, and your recall list is genuinely worked, the maths does not clear. You would be paying to solve a problem you already solved with people.
It is also wrong if the real problem is conversion rather than answering. A practice that answers every call and still cannot fill the book has a pricing or scripting problem, and adding a machine to the front of it will only fail faster.
Want to know if the numbers work for your practice?
Send how many calls you miss in a week, your average new-patient value, and roughly how many patients are on your recall list. We will tell you what it would recover — and say so plainly if it would not pay for itself.