Why a Third of Calls to Dental Practices Go Unanswered, and What AI Is Actually Fixing in 2026

The math no one wants to do

I ask the same question when I sit down with a four-chair dental practice. "How many calls do you think you're missing in a typical week?"

The answer is always the same: "A few. Maybe four or five."

Then we pull the actual call logs from the phone system. The number is usually closer to forty. Not a few. Not five. Forty per week.

This is not a failure of the front desk. It is a structural leak in how dental practices are built. Most offices have one or two people answering phones, and those people have other jobs — checking patients in, processing insurance, restocking, running the back. The phones ring during all of that. Between 20 and 35 percent of calls go unanswered in the average practice, according to industry reporting on AI dental receptionists from mConsent and Viva AI.

The practices that never run the math never feel the leak. The ones that do run the math usually find a revenue number that is, to put it politely, uncomfortable.

The specific times calls are lost

When I audit practice call logs, the missed calls are not evenly distributed. They cluster in four predictable windows. Knowing which windows is the first step to fixing the problem.

Window 1: The lunch drop (12:00 to 1:45 pm)

The front desk is eating. Someone is supposed to be covering. The cover is usually a dental assistant who is also sterilizing instruments. Calls ring into voicemail. The voicemail is often full.

Lunch alone typically accounts for 15 to 25 percent of weekly missed calls.

Window 2: The after-hours window (6 pm to 9 am, plus all weekend)

This is the big one. Most prospective patients decide to book care in the evening, after work, often while they are sitting on the couch thinking about a tooth that has been bothering them. They call the first practice on their list. If that practice sends them to voicemail, they call the second.

After-hours answering services built their entire business on this gap. The catch is that most of them are not really answering services anymore. They are forwarding services that take a message and email it to the practice. By the time someone at the practice sees the message, the patient has already booked with a competitor.

Window 3: The new-patient call during a busy morning

The phone rings at 9:47 am on a Tuesday. Everyone at the practice is with a patient. The phone drops to voicemail. The patient does not leave one. They move on.

This is the most painful category because these calls are not lost to schedule or to timing. They are lost to a simple staffing math problem: one or two humans cannot answer two or three simultaneous calls.

Window 4: The emergency that falls between chairs

"My crown just came off." "My kid fell and chipped a tooth." These calls are time-sensitive and high-value. If the front desk is mid-appointment, they often ring through to an out-of-office greeting. Some practices have emergency protocols. Most do not.

What dental marketing companies will not tell you about this

If you hire a dental marketing agency to "get more leads," they will sell you Google Ads, Meta Ads, SEO, maybe a new website. They will not tell you about the call leak, because it is not their product.

Here is the uncomfortable bit. If your practice is leaking 30 percent of calls, every marketing dollar you spend subsidizes that leak. You pay $85 to generate a call. One third of the time that call hits voicemail. You effectively paid $127 per answered call.

I have watched this dynamic kill marketing budgets at practices that thought they were buying leads. They were buying missed leads.

The cost in real dollars, for a real practice

Let me walk through the math on a four-chair general dental practice in a US metro. I am not making these numbers up. They come from actual client audits.

• Incoming calls per week: 130
• Missed call rate: 30 percent
• Missed calls per week: 39
• Missed calls per year: 2,028

Of those missed calls, industry averages say about 25 percent are new-patient inquiries. So 507 new-patient calls per year, lost.

The average new dental patient is worth roughly $750 in first-year revenue and significantly more over lifetime. Conservative estimate: if 20 percent of those missed new-patient calls would have converted to a booked first visit, that is 101 lost patients. At $750 first-year revenue, that is $76,000 in lost first-year revenue for one four-chair practice.

And this is before counting lifetime value, referral loss, or the opportunity cost of ad spend wasted on leads that hit voicemail.

The practice's marketing budget, if they spend $3,500 a month, is $42,000 a year. The call leak is larger than the marketing budget. No one tells them this.

What the old answering services got wrong

The traditional medical answering service industry — myreceptionist.com, abby.com, and similar — was built for a different era. They solved the "someone needs to answer the phone" problem by routing calls to humans in centralized call centers.

That worked in 2008. It works less well now for three reasons.

Reason 1: The humans are not in your practice. They cannot book into your scheduling software in real time. They take a message and forward it. By the time the practice receives the message, the lead has moved on.

Reason 2: The humans are not trained on your specifics. They do not know whether you take this insurance plan, whether the hygienist who handles sensitive teeth is Tuesday or Thursday, whether your implant consult is free or $150.

Reason 3: The cost model is per-call or per-minute. The busier the practice, the more the answering service costs. That creates a perverse incentive to stay small or to tolerate the leak.

These services are not bad. They are obsolete. And they know it. That is why you are starting to see them rebrand with "AI-powered" wrappers in 2026, which in most cases means a script more flexible than the one they had before, not a genuinely agentic AI receptionist.

What actually changed in 2026

Modern AI receptionists, the ones that work, do three things the legacy answering services cannot do.

They book in real time. The AI has direct API access to the practice management system. It sees the schedule, offers open slots, and confirms the booking during the call. The patient hangs up with a confirmed appointment and a text message.

They handle multilingual calls natively. Spanish, Portuguese, Mandarin, Vietnamese. The US dental market has a meaningful non-English-primary patient base. Most practice front desks handle one language. AI handles many.

They never get overwhelmed. If five calls come in simultaneously, five AI receptionists pick up. There is no queue. No hold time. The practice never sounds busy, because it structurally cannot be busy.

The shift from legacy answering to agentic AI is not marginal. In the Holistic Bio Spa case, flipping from human-only reception to a hybrid AI-plus-human stack took the practice from chronic call leak to 334 booked consultations in a single month. They did not increase their ad spend. They just stopped losing the calls they were already paying for.

Same pattern at Simetris: the 308x ROAS quarter was not driven by smarter ads. It was driven by the fact that every lead the ads generated was actually picked up, qualified, and booked, including the 7 pm Thursday calls that used to go to voicemail.

What to do if you suspect your practice is leaking

Here is the honest sequence. Skip the ones that do not apply.

Step 1: Run the actual numbers. Pull your call logs for the last 30 days. Count missed calls by hour. Not estimates. Actual data. If your phone system cannot export this, your phone system is the problem, and that is step zero.

Step 2: Categorize the misses. How many were after-hours? How many were lunch? How many were busy-signal during the work day? The fix is different for each.

Step 3: Calculate a revenue estimate. Use the math above. Missed calls times new-patient percentage times conversion rate times patient value. Do not let yourself round down.

Step 4: Decide the stack. Options: better human staffing (expensive, does not scale), legacy answering service (cheap, loses real-time booking), AI receptionist (upfront setup, compounds). Most practices end up at option three because the math is unambiguous once you run it.

Step 5: Run a 30-day pilot. Pick a single window — either after-hours or lunch coverage — and deploy the AI receptionist there only. Measure booked appointments from that window before and after. If the number does not move, the stack is wrong.

What we do at Hillflare, specifically

We build the stack that handles this on our medical AI page. The AI receptionist integrates with your practice management software (Dentrix, Open Dental, Eaglesoft, Nexhealth), handles English and Spanish natively, escalates clinical and billing calls to your human team, and attributes every booked appointment back to the ad, SEO page, or referral source that drove it. That last part matters because without attribution, you never know which marketing is actually working.

If you want to know whether this makes sense for your specific practice before buying anything, we offer a free growth diagnosis. It includes a call-audit review where we look at where your practice is actually leaking and tell you honestly whether AI is the right first lever.

You probably do not need more leads. You probably need to stop losing the ones you are paying for.

— Hector Arriola, Founder & CEO, Hillflare