Here is the uncomfortable thing about after-hours calls: you have no idea how many you get. Not because the data is hard to find — it is sitting in your phone system right now — but because a call that rings out leaves no trace anywhere you normally look.


Where the call actually goes

A patient decides at 7pm to sort out the tooth that has been bothering them for a week. They search, they tap the first number, and it rings. What happens in the next thirty seconds decides who gets the appointment.

If it goes to voicemail, most people do not leave one. They hang up and tap the next result. This is not a claim about your practice specifically — it is the behaviour of anyone with three more options on the screen in front of them. The call did not fail loudly. It just ended, and the patient became someone else's new patient.

The call that rings out is invisible in every report you look at. Your no-show rate does not move. Your production numbers do not move. Nothing in your practice management software has a field for "patient who tried to reach you and gave up."

Find out your own number — it takes ten minutes

Do not take anyone's industry average, including ours. Get your own figure:

  • Open your phone system's call log — VoIP providers, and most modern PBXs, keep inbound call records with timestamps and outcomes.
  • Filter to inbound calls outside your opening hours, plus any marked missed, abandoned, or unanswered during hours.
  • Count them over a full month, so you capture both the quiet weeks and the busy ones.
  • Cross-check against voicemails actually left in the same period. The gap between those two numbers is your answer.

That gap — calls received minus voicemails left — is the population of people who wanted to reach you badly enough to dial, and not badly enough to leave a message. Multiply it by the value of a new patient in your practice and you have the real annual cost, in your currency, with no vendor's estimate involved.

Why the missed calls cluster where they do

Three patterns show up in almost every practice's log:

  • Immediately after closing. The hour after you shut is often the busiest hour for inbound calls, because it is the first moment your patients are free to make them.
  • Your own peak hours. Mid-morning and mid-afternoon, when the front desk is checking someone in and the phone rings anyway. The second concurrent caller hears a busy tone or a queue, and queues lose people.
  • Monday morning. The weekend's worth of intent arrives at once, and the same one-call-at-a-time constraint applies.

Notice that only one of those three is an after-hours problem. Two of them happen while you are open and fully staffed — which is why "we'll just check voicemail more often" does not fix it.

The options, honestly

Extending staffed hours works and is expensive. A front desk hire in the US runs $2,500–3,500 a month plus employment costs, covers one shift, takes leave, and still answers one call at a time.

An answering service covers the hours and takes messages. It rarely books, because it usually cannot see your appointment book — so it converts the missed call into a callback task for tomorrow morning, which is better than nothing and worse than a booking. We compared the two properly in AI receptionist vs answering service.

An AI receptionist answers every call including the simultaneous ones, and — with scheduler integration — offers a real slot and books it before the caller hangs up. The patient goes to bed with an appointment and a confirmation text, and your front desk finds it already in the book.

Doing nothing is a legitimate choice if your log shows the gap is small. Run the ten-minute exercise first. If you are missing four calls a month, none of this is worth your attention. If you are missing sixty, it is the highest-leverage thing on your list.

What good after-hours handling looks like

  • It answers, fast. Two rings, not eight. Delay reads as absence.
  • It is honest. The caller is told they are speaking to an AI assistant, at the start. Nobody likes discovering it late.
  • It can say yes to a time. Access to live availability is the difference between a booking and a message.
  • It knows what it must not touch. Clinical questions route to a human, every time, by rule — not by the model's judgement in the moment.
  • It escalates urgency properly. Pain and trauma language goes to your on-call protocol immediately, not into a booking flow.
  • It confirms in writing. An SMS confirmation turns a late-night call into an appointment the patient will actually remember.

That is the specification we build to. If you want the details, the AI receptionist page covers how it works and pricing is published in full. But run your own call log first — the number you find there should decide this, not an article.