Every dental practice has a growth plan. Very few have an answer to the simplest question in the business: what happens when someone rings at 11:40am while the front desk is checking in a patient, taking payment from another, and holding a line for the lab?

The answer, usually, is nothing. The phone rings out.

The scale of it is larger than most owners think

Call analytics firm Peerlogic analysed 4,280 inbound calls across 26 dental practices in February 2026 and found that 38% went unanswered. Their wider reporting puts the answer rate for new patient calls at around 68%, and of the calls that are answered, only about 42% convert into a booked appointment.

These figures come from a company that sells call handling technology, so treat them as directionally useful rather than neutral. But the range they describe, somewhere around a third of calls missed, matches what practice owners tell us when they actually go and pull their own phone records. That is the exercise worth doing before believing anyone, including us. Your phone system almost certainly holds this data already.

The reason missed calls matter more in dentistry than in most industries is what the caller does next. A prospective patient who reaches voicemail very rarely leaves a message. They ring the next practice on the list. There is no second chance and, importantly, no record. The loss is completely invisible in your practice management software, which is why it survives for years in businesses that are otherwise carefully managed.

You are not looking at a lost appointment. You are looking at a lost patient relationship, which in dentistry is measured in years of routine care plus whatever treatment they eventually need.

The second leak: people who booked and did not arrive

The other side of the same problem is the appointment that exists but goes unfilled.

There is good evidence here rather than vendor claims. A randomised controlled trial published in the American Journal of Medicine in 2010 assigned just over 10,000 patients at an outpatient practice to three groups: a reminder call from clinic staff, an automated telephone reminder, or no reminder at all. Reminders went out three days before the appointment.

The no-show rates were 23.1% with no reminder, 17.3% with an automated reminder, and 13.6% with a staff call.

Two things are worth pulling out of that, and the second one is the part vendors leave out.

First, automated reminders clearly work. Cutting no-shows from 23.1% to 17.3% is a meaningful reduction, and it happens without consuming any staff time at all.

Second, the human call performed better than the automation. Significantly so. Anyone telling you that automation strictly outperforms a person on this specific task is selling something.

The correct conclusion is not that you should have humans do reminders. It is that automation should handle the volume so that your team has time to make the calls that genuinely benefit from a human voice, which is a very different design decision from replacing the front desk.

What a voice AI receptionist actually does well

Set expectations correctly and this technology is genuinely good. Set them wrong and you will annoy patients, which is expensive.

What it handles reliably:

  • Answering every call, including the ones outside opening hours. A significant share of calls to dental practices arrive when the practice is closed, from people who cannot ring during their own working day.
  • Overflow when the desk is busy. This is the highest-value use and the least disruptive. The team stays in place. Nothing rings out.
  • Booking, rescheduling and cancelling routine appointments directly against live availability.
  • Answering the questions asked forty times a week. Opening hours, parking, what to bring, whether a particular treatment is offered, roughly what it costs.
  • Reminder and confirmation sequences, including filling a slot that has just been cancelled by working through a waiting list automatically.

What it does not handle, and should not be asked to:

  • A patient in pain. These calls need a person immediately. Any sensible configuration routes them straight through.
  • Complaints or anything emotionally loaded. Escalate on the first signal.
  • Clinical questions. The line is straightforward. Logistics can be automated. Anything clinical cannot.
  • Complex treatment plan discussions, where the conversation is part of the patient making a decision.

A well-built system knows which of these it is in within a few seconds and hands over cleanly. A badly built one traps a patient with toothache in a menu, and you will hear about it.

Why these projects fail when they fail

The practices that get poor results from this almost always made the same mistake, and it is not a technology mistake.

They bought a phone AI and pointed it at the existing front desk process without ever describing what that process is. Nobody wrote down what actually happens when a new patient rings, which questions get asked in which order, how a slot is really chosen, what the exceptions are, or who gets consulted before a booking is confirmed. So the system was configured against an idealised version of the workflow, and every real-world variation became a failure the team then had to clean up.

Within a month the front desk trusted it less than the answering machine.

The fix is unglamorous and it comes first. Map the call flow as it genuinely operates, including the messy parts. Decide explicitly which call types are automated, which are escalated and where the boundary sits. Then configure. Practices that do it in that order tend to keep the system running. Practices that do it the other way round have usually switched it off by month three.

A regional note

We worked with a clinic in Riyadh that was losing SAR 11,000 a month to manual appointment handling. Missed calls, no-shows, double bookings and the staff time spent unpicking all three. Nobody in the practice had put a figure on it before, because none of it appeared as a line item anywhere. It was simply absorbed as the normal cost of a busy front desk.

That is the pattern. The number is rarely dramatic on any single day. It is dramatic annually.

Work out your own number first

Before you evaluate a single product, pull three figures from systems you already have. How many calls came in last month and how many were answered. What percentage of booked appointments were not attended. And how many of your answered calls actually resulted in a booking.

Those three numbers will tell you whether this is worth your attention, and they cost nothing to obtain.

If you would rather have someone do that properly, that is what our AI Audit covers. Over 14 days we map how work actually moves through the practice, put a real monthly cost on where it leaks, and hand back the three to five automation opportunities worth doing, ranked by impact, effort and readiness. It is $1,500, or SAR 5,600, and the roadmap is yours regardless of who builds it.

If the answer is that your front desk is already handling the volume well, we will say so. That is a useful thing to know for the price of finding out.