Voicemail isn't a safety net. 78% of after-hours callers who hit it never leave a message.

Your second site runs itself now. The third is the one you are actually weighing, and every spreadsheet built for it assumes the phones get answered the way they do today. They probably do not. A February 2026 analysis of 4,280 calls across 26 dental practices found that 38% went unanswered during normal business hours, and among the calls that were answered, only 42% converted into a booked appointment (Peerlogic dental call center case study, February 2026). That is before anyone accounts for what happens outside business hours, when the numbers get worse.

"If we're slammed, it goes to voicemail. Patients who actually want an appointment call back." That is the working assumption behind most front-desk staffing decisions, and it has a kernel of truth in it. It is also the belief that is quietly costing expanding practices the most money, because it treats a missed call and a lost patient as two different problems, when the data says they are almost the same problem.

The part of the myth that is true

No front desk answers every call. A two-clinician site with one person on reception cannot take a call while she is checking in a patient, verifying insurance for another, and confirming tomorrow's schedule. Adding headcount to close that gap is expensive and, per this site's own reporting on dental staffing, not always possible even when the budget exists: 39.7% of US dental practices report inadequate hygiene staffing as of Q1 2026, with 90.5% of dentists who tried calling it very or extremely difficult to hire for (ADA Health Policy Institute, Q1 2026). Front-desk roles compete for the same tight labor pool. Treating 100% live-answer as the target is not realistic, and nobody should build a plan around it.

So what this means for your practice: the honest goal is not zero missed calls. It is making sure a missed call does not automatically become a lost patient, which is a different and far more solvable problem than hiring your way out of a staffing shortage.

What voicemail actually catches

This is where the "they'll call back" assumption breaks down. Dental Intelligence, whose AI Receptionist add-on rolled out broadly to its dental customer base in mid-2026, states in its own July 2026 product documentation that nearly half of all inbound calls to a dental practice arrive after business hours, and that 78% of those after-hours callers do not leave a voicemail (Dental Intelligence, July 2026 product update, vendor-published). They simply hang up. During business hours the pattern is less extreme but still steep: of the 38% of calls that went unanswered in the Peerlogic dataset, only a minority converted through a callback or voicemail follow-up; most of the lost opportunity in that study came from the 62% of calls that were answered live but still did not book, at a 42% conversion rate.

Put those two findings together and voicemail is not functioning as a safety net for either half of your call volume. During the day, most of the leak is calls that get answered but mishandled. At night and on weekends, when almost half your total call volume is arriving, the leak is a caller who never speaks to anyone and rarely leaves a trace.

So what this means for your practice: if you have never pulled a report on what percentage of your after-hours calls end in a hang-up versus a voicemail, that number is very likely worse than you assume, and it is the single most measurable gap in your current setup.

What the ratio is actually costing

Patient Prism, a dental call-tracking analytics company, puts the value of a single missed call at $200 to $300 in immediate revenue and more than $15,000 in lifetime patient value once referrals and repeat visits are factored in (Patient Prism, 2026 healthcare call-tracking metrics report, vendor-sourced, not independently verified). Take that figure with real caution; it is published by a company that sells the fix. But it does not need to be exactly right to matter. A general dentist's average production runs $475 to $575 an hour, against average annual gross billings of $942,290 per GP dentist (ADA Survey of Dental Practice, 2024, an independent trade-association benchmark, not a vendor estimate). A two-site practice fielding a modest 60 calls a day, in line with industry-reported call volumes for small and mid-sized dental groups, is missing roughly 23 of them during business hours alone at a 38% miss rate, before after-hours volume is even counted. Even a fraction of those representing genuine new-patient inquiries that go elsewhere is real chair time, not a rounding error on the spreadsheet.

So what this means for your practice: before adding a third location to a call-handling system that already loses a third of its business-hours volume and most of its after-hours volume, run the arithmetic on your own call count. The third site multiplies the leak; it does not shrink it.

What actually closes the gap

The fix is not "train the front desk harder," although training helps at the margin. It is treating call handling as a system with a measurable answer rate and conversion rate, the same way you already treat production and collections. Several 2026 platforms are built specifically to do this: Dental Intelligence's AI Receptionist answers overflow and after-hours calls, verifies patient identity, and hands staff a summary the next morning instead of a stack of voicemails (Dental Intelligence, July 2026). Peerlogic's Voice AI and text-back tools follow a similar model, automatically texting a missed caller within seconds so the conversation continues even when nobody picked up. Among the tools worth evaluating, the practical differences are in PMS integration depth, after-hours coverage, and price, not in whether the underlying idea works; the vendors themselves report meaningful recovered revenue from customers who adopt these tools, though those figures are self-measured and should be verified against your own call data rather than taken at face value.

One point of diligence matters regardless of which platform you evaluate: any AI tool that answers a phone, confirms a patient's identity, or discusses an outstanding balance is handling protected health information, and under HIPAA that makes the vendor a business associate. A signed Business Associate Agreement covering the AI layer specifically, not just the practice management system it plugs into, is non-negotiable before deployment (HHS Office for Civil Rights, HIPAA Business Associate guidance).

So what this means for your practice: the evaluation question is not "does AI answering work," the market has already answered that. It is "what is my current answer rate and conversion rate, and does closing that gap justify the monthly cost," which is a number you can calculate in an afternoon with two weeks of call logs.

The call

The evidence does not support treating voicemail as a functioning backup for the calls your practice cannot answer live. It functions for a small share of business-hours misses and close to none of your after-hours volume. Before opening a third location, or before spending on any AI call-handling tool, pull two weeks of call logs and measure your own answer rate and conversion rate. If they land anywhere near the 38% miss rate and 42% answered-call conversion this data points to, the math for closing that gap with a dedicated system, rather than more headcount you may not be able to hire anyway, is difficult to argue against.

For a scored, independent read on where your call handling actually stands before you commit budget to a fix, or open a third site on top of an unmeasured problem, the AI Opportunity and Growth Assessment covers exactly that comparison. Start with a free 20-minute discovery call.

See also: why AI recall and call-handling tools work for independent practices, not just DSOs, and what the same 2026 staffing squeeze means for hygiene capacity.

Related: 84% of US dentists are independent, and AI recall works there too · The dental staffing shortage AI can and cannot fix · Dental AI already standardizes across 1,900 locations