How Best Smiles Dental Generated ~$1.5M in New-Patient Revenue Across 5 Clinics with AI

Best Smiles Dental (BSD) is a five-location dental group whose growth was bottlenecked by the place every dental practice quietly loses revenue: the front desk. Across SouthSide, StaplesMill, GlenAllen, NorthSide, and Dr. C Advanced Dental, each office ran its own phones, its own schedule, and its own coverage gaps. When call volume spiked, lines went to hold or voicemail. When the office closed, prospective patients reached an answering machine and often the next practice's website. Ownership had five separate views of performance and no single, comparable picture of how many opportunities were arriving or being missed.

In May 2026, BSD replaced that fragmented model with TensorLinks AI, a centralized AI front desk built specifically for multi-location dental groups. A single AI receptionist now answers every inbound call around the clock with zero hold time and no voicemail, handles scheduling across voice, text, and online channels, qualifies new patients, and reschedules or routes callers as needed. Each clinic kept its own schedule and routines. Ownership gained one command center spanning all five offices. The design goal was straightforward: enterprise visibility paired with local efficiency.

Month-One Headline Results

Over a single month, the centralized AI front desk handled the entire inbound load and turned it into measurable bookings:

• 2,366 inbound calls answered across the five locations, with roughly 3,993 minutes of AI talk time (about 66.5 hours).
• 163 appointments created and 350 appointments managed in total, including reschedules and routing.
• 129 new patients onboarded in the month.
• 6.9% blended call-to-booking conversion, with the strongest locations clearing 8.9% and 9.1%.

Translating those new patients into revenue requires an assumption, so we state it plainly. BSD chose to value each new patient at approximately $1,000 in first-year production, a figure within commonly cited ranges for general dentistry. On that basis, 129 new patients in one month represents an estimated $129K in first-year value, or roughly $1.5M annualized across the group if the month-one pace holds. This is an illustrative estimate, not booked revenue, and actual production varies by case mix, payer, and treatment acceptance.

The Central Thesis

The result that matters here is not any single metric. It is the shift in what the front desk became. For most dental groups, reception is a cost center that caps growth: limited by staffing, business hours, and the human ceiling on how many calls one person can answer well. Consolidating five separate front desks into one AI-driven layer changed that equation. Every call got answered, after-hours demand stopped leaking away, and every interaction was captured as data ownership could see and act on.

When reception stops being a capacity constraint and starts being measurable, the front desk turns from a cost center into a growth engine.

That visibility also surfaced where the upside still sits. NorthSide converted at 2.9% and StaplesMill at 3.8%, well below the top performers, even though both fielded meaningful call volume. The same system that lifted the group average exposed a clear, addressable conversion gap, which is itself part of the return.

What This Study Covers

The sections that follow trace the full arc. We profile BSD as a five-location group and the economics of the dental front desk, then frame the specific challenge of running five reception teams against one growth ceiling and why centralized AI beat additional hiring. We walk through how the TensorLinks AI front desk works and how it went from kickoff to live. From there we move into the numbers: a complete month-one breakdown, deep dives into the high performers and the three locations with room to grow, the revenue model behind the $1.5M estimate, and a focused look at closing the conversion gap. We close with capacity and after-hours impact, operational and patient-experience effects, the visibility-versus-efficiency balance, and practical lessons, an FAQ, and the path forward for other multi-location groups.

From the outside, a five-location dental group reads as one brand. Patients see a single name on the sign, a shared website, maybe a unified phone number on the billboard off the interstate. Inside, the reality is closer to five small businesses that happen to share an owner. Best Smiles Dental runs SouthSide, StaplesMill, GlenAllen, NorthSide, and the affiliated Dr. C Advanced Dental as distinct operations, each with its own front desk, its own phone lines, its own daily rhythm, and its own relationship with the patients who live nearby.

That structure is not an accident. It is how dental groups grow. A practice does well, the owner acquires or opens a second location, then a third. Each addition arrives with its existing staff, its installed phone system, and a book of patients who think of "their" office as the one down the street, not as branch four of a regional group. The group scales by stacking these units, and the front desk scales the same way: one more office means one more set of phones and one more team to answer them.

Five Front Desks, Five Realities

The first thing that becomes clear when you look across BSD's five locations is how uneven the load is. In May 2026, inbound call volume ranged from 809 calls at SouthSide down to 276 at NorthSide. SouthSide alone handled roughly a third of the group's 2,366 total inbound calls. StaplesMill took 548, Dr. C Advanced Dental 395, and GlenAllen 338. The busiest office fielded nearly three times the call traffic of the quietest one.

That spread matters because each front desk is staffed against its own location's needs, not the group's average. A high-volume office like SouthSide can have every line ringing at once during the morning rush while a coordinator is checking in a patient, processing a copay, and answering a clinical question for the hygienist. A lower-volume office like NorthSide may sit quiet for stretches, then get slammed when two staff are at lunch. Neither office can lend capacity to the other. The phones do not pool. When SouthSide's lines are full, the eleventh caller does not get routed to a coordinator sitting idle across town. They get a busy signal, a hold, or voicemail.

Ownership feels this as a visibility gap more than a staffing one. The owner of a five-location group rarely knows, on any given Tuesday, how many calls each office missed, how long callers waited, or how many would-be new patients hung up before booking. The data lives in five separate phone systems and five separate front-desk memories. By the time it surfaces, usually as a slow month or a frustrated review, the patients are already gone.

Why Linear Scaling Breaks

The instinct, when a group grows, is to scale reception in lockstep with locations. Open a sixth office, hire a sixth front desk. Add Saturday hours, add a coordinator. On paper this is tidy. In practice it runs into three structural problems that compound as the group gets bigger.

• Capacity is trapped per location. Each front desk can only answer the calls coming into its own building. There is no shared queue, so peaks at a busy office and lulls at a quiet one cannot offset each other. The group pays for enough staff to cover every location's peak, then watches that capacity sit idle the rest of the day.
• Coverage gaps multiply. A single office has predictable holes: lunch, end of day, the moment two patients arrive at once, evenings and weekends when the phones are off entirely. Across five offices, those gaps overlap into a wide window where calls go unanswered somewhere in the group at almost any hour.
• The front desk is doing two jobs. A coordinator standing at the desk is responsible for the patient physically in front of them and the patient on the phone. When the lobby is busy, the phone loses. Every group with a front desk has quietly accepted that some percentage of inbound calls will never be answered well, because the person answering them is also running the office.

The result is a growth ceiling that has nothing to do with clinical capacity. BSD's operatories could see more patients. The dentists have open chair time. The constraint sits at the front of the building, in the gap between calls coming in and calls being converted into booked appointments. Across the group, only 6.9 percent of inbound calls became a created appointment in May, and that blended figure hides a wide range, from 9.1 percent at the strongest office to 2.9 percent at the weakest.

The bottleneck in a growing dental group is rarely the dentist's chair. It is the distance between a call coming in and a patient getting booked, repeated across every location at once.

This is the structural problem the rest of this study works through. A five-location group is not one front desk scaled up; it is five front desks that cannot help each other, each managing its own ringing phones, its own coverage gaps, and its own conversion rate, with ownership watching from a distance through fragmented data. Adding a sixth office adds a sixth version of the same problem. The question BSD faced is the one any multi-location group eventually reaches: how do you give every location more reception capacity without simply hiring more people at every location, and how do you finally see the whole group at once. The sections that follow trace how a centralized AI front desk answered both.

Every dental practice already owns a growth channel more valuable than any ad campaign: the phone line. When someone calls a dental office, they are rarely browsing. They have a toothache, a cracked filling, a new insurance card, or a child who needs a first cleaning. The call itself is a qualified lead with intent, and the only question is whether the practice converts it. That is why the front desk, often treated as an administrative cost, is more accurately understood as a revenue function. The economics are simple to state and easy to underestimate.

An inbound call is a unit of potential revenue

Treat each inbound call as a small option on future production. Some callers want to confirm an appointment or ask a billing question, but a meaningful share are trying to book, reschedule, or become new patients. Industry estimates commonly place the first-year production value of a new dental patient somewhere in the four-figure range once you account for the exam, imaging, hygiene, and the restorative or elective work that often follows. Best Smiles Dental chose a conservative round number of roughly $1,000 per new patient for first-year production. We use that figure throughout this case study as an illustrative estimate, not a guaranteed result, and your own number will depend on your case mix and fee schedule.

Run the arithmetic on a single location. If a practice fields a few hundred calls a month and converts even a single-digit percentage of them into booked new patients, the revenue attached to that phone line dwarfs the salary of the person answering it. The leverage is not in handling calls cheaply. It is in not losing the ones that carry intent.

The quiet drain: missed calls and voicemail

Here is where the math turns against most practices. A front desk staffed by people is, by definition, sometimes busy. Someone is checking out a patient, processing a claim, or already on another line. Calls that arrive during those moments ring out, hit voicemail, or get a busy signal. The patient, who had intent a second ago, simply dials the next office on their list.

Industry observers have long flagged that a surprising fraction of inbound calls to dental and medical offices go unanswered during business hours, with estimates frequently cited in the range of one in four to one in three depending on the practice and the day. Add the calls that arrive after hours, on weekends, and during lunch, and the unanswered share climbs further. The conversion economics of those two outcomes are not close.

Voicemail looks like a safety net, but for new-patient acquisition it functions closer to a dead end. A prospective patient comparing offices is not waiting by the phone for a return call. By the time a staffer clears the queue and dials back, the appointment is often booked somewhere else. Every missed call is therefore not a deferred conversion. It is usually a lost one, and the loss is invisible because it never enters the schedule, the report, or anyone's awareness.

Acquisition cost versus the cost of a dropped call

Most groups spend real money to make the phone ring: paid search, directory listings, referral programs, and reputation management all carry a cost per lead. When marketing dollars generate a call that then goes to voicemail, the practice pays twice. It pays the acquisition cost to create the inquiry, and it forfeits the first-year value of the patient that inquiry could have become. On an estimated $1,000 first-year basis, a handful of dropped new-patient calls each week quietly erases the return on an entire marketing budget. The front desk, in other words, is where marketing spend is either converted or wasted.

The cheapest patient to acquire is the one already calling you. The most expensive mistake is letting that call go unanswered.

Why the phone is the highest-leverage surface in dentistry

Several factors stack up to make the phone the single most concentrated point of growth in a dental business. Demand arrives with intent rather than needing to be manufactured. The cost of capturing that demand is mostly a matter of availability, not additional advertising. And the downside of failure is silent, so it rarely gets fixed. A practice can run a flawless clinical operation and still cap its own growth at the front desk without ever seeing the ceiling.

This is the frame that makes the Best Smiles Dental results worth examining closely. Across five locations in May 2026, the group recorded a 6.9% blended call-to-booking conversion and onboarded 129 new patients in a single month. On a stand-alone basis those figures are just numbers. Read through the economics above, they are something more concrete: 129 captured units of estimated first-year production that, in a voicemail-and-busy-signal world, a portion of which would have gone uncounted and unbooked. The conversion rate is not a vanity metric. It is the percentage of revenue-bearing intent the group managed to keep.

The sections that follow put those numbers under a microscope, location by location, and show how the spread between the group's best-converting and worst-converting clinics translates directly into dollars left on or taken off the table. But the principle holds before a single location is named. In dentistry, the phone is where growth is won or quietly surrendered, and the difference between the two is a matter of whether every call gets answered.

On paper, Best Smiles Dental was running five healthy practices. SouthSide, StaplesMill, Dr. C Advanced Dental, GlenAllen, and NorthSide each had their own front desk, their own phones, and their own rhythm. Yet anyone who has owned a multi-location group knows the feeling that the parts were adding up to less than they should. Demand was clearly there. In a single month the five clinics fielded 2,366 inbound calls. The question ownership could not answer with confidence was simple: how much of that demand was actually turning into patients, and where was it leaking away?

That uncertainty is the real challenge. It is not that any one location was failing. It is that nobody had a single, trustworthy view of how the group converted phone calls into booked chairs. The symptoms were felt long before they could be measured.

One group, five different conversion rates

The clearest signal in BSD's own numbers is how widely the locations diverged on the metric that matters most: turning an inbound call into a booked appointment. Across the group the blended call-to-booking conversion was 6.9%. But that average hid a spread that should make any owner uncomfortable.

The top performer converted callers more than three times as effectively as the bottom one, 9.1% against 2.9%. These are practices under the same brand, with comparable services and a shared standard of care. A gap that wide is not about patient quality or local market. It points to something procedural at the moment a phone rings: who answers, how fast, what they say, and whether the call ends with a date on the calendar.

Consider what the variance implies in raw terms. StaplesMill handled 548 calls, more than Dr. C's 395, yet booked far less of that traffic. If StaplesMill had simply matched the conversion of its higher-performing siblings, its appointment output would have looked entirely different from the same volume of demand. The traffic was never the constraint. The handling was.

No portfolio-wide view to act on

Here is why the gap persisted: ownership had no single place to see it. Each front desk reported in its own way, on its own cadence, if at all. There was no command center that lined the five locations side by side and said, in effect, "NorthSide is converting at less than a third of Dr. C's rate this month." Without that view, a 2.9% location and a 9.1% location look the same from the top. Both are "busy." Both have ringing phones and full waiting rooms at peak. The difference only surfaces when you can compare them on the same axis at the same time, and BSD could not.

This is the quiet tax of running locations as islands. Best practices stay trapped inside the clinic that discovered them. Whatever Dr. C's team was doing to reach 9.1% never traveled to NorthSide, because no one had the data to know it was worth copying. Problems and solutions both went invisible.

A multi-location group does not have a performance problem until it can measure one. The danger of running each front desk as an island is that a three-to-one conversion gap can hide in plain sight, quarter after quarter.

After-hours and overflow: the demand you never hear

The 2,366 figure counts the calls BSD answered. It says nothing about the ones that arrived after closing, during lunch, or while every line was tied up at the morning rush. A traditional front desk is only open when staff are at the desk, which means a meaningful share of inbound interest hits voicemail or a busy signal and is gone. Industry estimates commonly put the share of missed or unreturned calls at dental practices in the double digits, and voicemail conversion is typically poor. We are not assigning BSD a specific miss rate, because the legacy setup could not capture one. That is precisely the point: the calls that never reached a person were also the calls that never reached a report.

For a group carrying this much volume, even a modest after-hours and overflow leak compounds across five locations and twelve months. Every one of those callers was a prospective patient deciding, in real time, whether Best Smiles was reachable.

Inconsistent new-patient capture

The same unevenness shows up in new-patient onboarding. The group brought on 129 new patients in the month, but the distribution mirrored the conversion spread: SouthSide added 55 while NorthSide added 8. New patients are the engine of practice growth and the most valuable callers a front desk handles. When the process for qualifying and booking them depends on which clinic and which staff member happens to pick up, capture becomes a matter of chance rather than design.

The growth ceiling, stated plainly

Put the symptoms together and BSD's situation comes into focus. Demand was strong and growing. Handling was inconsistent by location, invisible from the top, leaky after hours, and unreliable for the highest-value callers. None of these were dramatic failures. They were structural limits baked into the model of five separate front desks, and they capped what the group could grow into no matter how much marketing fed the phones. The next sections look at why adding more staff would not have lifted that ceiling, and what did.

By the time a dental group reaches five locations, the front desk stops being a staffing question and becomes an architecture question. Every clinic answers its own phones, keeps its own routines, and absorbs its own call volume. When demand spikes or a receptionist is out, the easy reflex is to add another person. But more hiring is only one of four ways to handle the phone, and it is rarely the one that fixes the underlying problem. Below is how the realistic options compare against the things that actually moved Best Smiles Dental's numbers in May 2026: cost behavior, consistency, around-the-clock coverage, conversion quality, and how well the approach holds up across multiple sites.

The four options on the table

Most multi-location groups choose among hiring more front-desk staff, contracting a traditional human answering service, deploying an IVR or phone tree, or moving to a centralized AI front desk. Each solves part of the problem and creates a different set of constraints.

Why more hiring stalls out

Hiring is the most familiar lever and the one with the worst economics at scale. Front-desk roles carry recruiting time, training time, benefits, turnover, and the quieter cost of a half-trained hire converting fewer callers for the first few months. The deeper issue is that a person can only be on one call at a time. The moment two lines ring together, one caller hits hold or rolls to voicemail, and industry estimates commonly put the share of prospective patients who simply hang up on voicemail at a large fraction. A receptionist also cannot work the 7 p.m. call from someone in pain or the Saturday inquiry from a family shopping for a new dentist. Adding staff raises the ceiling a little and raises fixed cost a lot, and it does nothing for the hours when no one is at the desk.

Why answering services and phone trees fall short

A traditional human answering service buys coverage but usually not conversion. These services typically take messages and pass them back to the clinic, which means a motivated caller is contacted again later, if at all. Every handoff is a chance to lose the patient. Cost also climbs with volume, so the busiest months, exactly when the most new patients are calling, are the most expensive.

IVR and phone trees are cheap and tireless, and that is roughly where their advantages end. They are built to route, not to convert. New patients who are deciding between practices do not want to press four to leave a message; many abandon the call. For a group trying to grow, a phone tree optimizes for cost per call while quietly suppressing the metric that matters, which is whether the call turns into a booked appointment.

What centralized AI changes

A centralized AI front desk is the only one of the four options that pairs always-on coverage with the ability to qualify and book in the same conversation. It answers every inbound call with zero hold and no voicemail, handles several calls at once, applies the same qualification logic on the first ring and the last, and never has an off day. Cost tracks usage rather than headcount, so capacity rises with demand instead of requiring a new hire ahead of it.

The data from BSD's first month reflects this. Across the five clinics the system handled 2,366 inbound calls and roughly 66.5 hours of AI talk time, created 163 appointments, and onboarded 129 new patients at a blended 6.9% call-to-booking conversion. Those new patients arrived without adding a single front-desk position. At the client's chosen and explicitly illustrative figure of about $1,000 in first-year production per new patient, that volume implies roughly $129K in a month, an estimate meant to frame scale rather than promise a return.

The most expensive call a growing practice handles is the one no one answers. Coverage without conversion is just a more polite way to lose a patient.

The point is not that AI replaces a great front-desk team. It absorbs the overflow, the after-hours volume, and the repetitive qualification that pulls staff away from patients standing at the counter. The fifth option also does something the other three cannot: it gives ownership one view across every location while each clinic keeps its own schedule. That combination of enterprise visibility and local efficiency is the reason a five-location group reaches for centralized AI instead of a sixth round of hiring.

The fix for Best Smiles Dental was not a better phone tree or a bigger call center. It was a single AI front desk that sits across all five locations at once, answers every inbound call, and books patients straight into the right schedule. The same agent that handled SouthSide's 809 May calls also handled NorthSide's 276, without either clinic giving up its own rules, providers, or routines. What follows is how that works at a capability level.

One agent that answers every call, 24/7

At the center is an AI receptionist that picks up on the first ring, around the clock. There is no hold music, no voicemail, and no after-hours gap. A patient calling at 7:00 a.m. on a Saturday gets the same greeting and the same booking ability as one calling at 2:00 p.m. on a Tuesday. Across the group, that consistency added up to 2,366 answered calls and roughly 66.5 hours of AI talk time in a single month, all of it handled without adding a person to any clinic's payroll.

Because the agent is software, it does not queue. Five calls arriving at SouthSide in the same minute are five simultaneous conversations, not four callers waiting while one is helped. That removes the structural reason most dental front desks lose calls: a finite number of humans answering a phone that rings in bursts.

Unified scheduling across voice, online, and text

Patients reach BSD through more than one door, and the AI front desk treats all of them as the same conversation. Whether someone calls, fills out an online request, or sends a text, the agent can find an open slot and book it. The patient does not need to know which channel maps to which staff member. They state what they need, and the system finds time that fits.

This matters most for new patients, who rarely behave predictably. One starts a request on the website at midnight, abandons it, then calls the next morning. Another texts a question about availability and books on the spot. Unifying voice, online, and text under one agent means none of those threads dead-ends at a channel no one is watching. In May, that unified intake produced 163 appointments created and 129 new patients onboarded across the group.

Automatic new-patient qualification

Booking a new patient is more than dropping a name in an open slot. The visit type has to be right, the basic information has to be captured, and the appointment has to land where the practice can actually serve it. The AI front desk handles that qualification in the same conversation: it identifies whether the caller is new or returning, gathers what the clinic needs to prepare, and books the appropriate appointment length and type.

Done consistently, qualification protects the schedule. A new patient routed to the correct visit type is less likely to arrive for the wrong service, and the front office is not left reconstructing details after the fact. For a group adding more than a hundred new patients a month, that upfront accuracy compounds.

Rescheduling and routing without a callback

Schedules change. The agent handles reschedules and cancellations in real time rather than collecting a message for someone to return later. When a request is outside what the AI should resolve on its own, it routes the caller to the right place at the clinic rather than dropping them into a generic queue. The 350 appointments managed across the five locations in May include this ongoing work of moving, confirming, and rerouting appointments, not just the initial bookings.

The practical effect is fewer open loops. A patient who needs to move an appointment does it in the moment. The clinic does not accumulate a stack of voicemails to work through, and the schedule reflects reality instead of yesterday's intentions.

Local rules, honored per clinic

A centralized agent only works if it does not flatten five practices into one. Each BSD location keeps its own schedule, its own providers, and its own way of running the day. Dr. C Advanced Dental does not operate identically to GlenAllen, and the AI front desk does not force it to. The agent applies each clinic's booking logic to that clinic's calls, so a patient calling StaplesMill is booked under StaplesMill's availability and rules, not a shared template.

This is the difference between centralization and standardization. The intelligence is shared. The execution stays local. One agent serves all five front desks while each clinic experiences it as its own receptionist who happens to never miss a call.

Centralizing the front desk does not have to mean standardizing the practice. The smarter pattern is shared intelligence with local execution, so every location gets the same coverage while keeping the routines that make it work.

The portfolio command center

For ownership, the value is not only that every call is answered but that the whole group is finally visible in one place. The command center gives leadership a single view across all five locations: call volume, appointments created and managed, new patients, and conversion, side by side. That is how it becomes obvious that SouthSide and Dr. C convert near 9 percent while NorthSide sits at 2.9 percent, a gap that is hard to even see when each clinic reports separately.

This is what the positioning, enterprise visibility meets local efficiency, means in practice. Each clinic keeps its own schedule and runs its own day. Ownership gets a portfolio-level read on performance without calling five offices and reconciling five spreadsheets. The same system that answers a 7:00 a.m. call in one location tells the operator, that night, exactly how every location performed.

Taken together, these capabilities are not five separate tools bolted onto five practices. They are one front desk that scales with call volume, works every hour, and reports up to ownership while serving each clinic on its own terms. The month-one numbers in the next sections show what that produced once it went live.

Replacing a front desk process across five clinics is less a software install than a sequence of carefully ordered decisions. The goal of onboarding is not to flip a switch on day one. It is to teach a single AI receptionist how five different locations actually run, prove it works on real calls, and only then hand it the full phone load. The phased approach below is the playbook a multi-location dental group can expect to follow. It mirrors how a deployment like Best Smiles Dental's reached the May 2026 results documented elsewhere in this study: 2,366 calls handled and 129 new patients onboarded in a single month.

Phase 1: Discovery and PMS connection

Everything starts with the practice management system. The AI front desk has to read live availability, write new appointments back, and look up existing patients, so the first technical step is establishing a secure connection to the group's PMS. For a five-location group this often means confirming whether all clinics share one database or run separate instances, mapping each location to its own schedule, and verifying that operatories, providers, and appointment types come through cleanly.

Alongside the connection, discovery captures the operational reality of each site. That includes office hours, the provider roster, the services each location offers, accepted insurance, new-patient policies, and any quirks one clinic has that another does not. The integration is validated in a sandbox before anything touches production scheduling. Nothing about the patient phone experience changes yet.

Phase 2: Training the AI on all five locations

With data flowing, the AI is trained location by location. Hours, providers, and services differ across sites, and the system must answer as though it belongs to whichever clinic the caller dialed. A patient asking the NorthSide number about Saturday hours should hear NorthSide's answer, not GlenAllen's.

Training also covers the conversations that fill a real day:

• New-patient qualification: confirming the reason for the visit, insurance, and contact details so the booked appointment is ready to be seen.
• Scheduling and rescheduling logic, including which appointment types map to which providers and how far out the AI may book.
• Frequently asked questions about location, parking, services, and payment.
• Escalation rules for clinical questions, emergencies, and anything that belongs with a human.

This is also where the group decides how unified scheduling should behave. The command-center model keeps each clinic's schedule and routines intact while giving ownership one view across all five, so training reinforces local accuracy rather than flattening every site into a generic script.

Phase 3: Routing, escalation, and scheduling rules

Configuration is where policy becomes behavior. The group defines how inbound calls route, what the AI handles end to end, and what gets transferred. Practical rules include guardrails on double-booking, handling of recall and existing-patient requests, after-hours behavior, and the path for callers who simply prefer a person.

A short rule-tuning pass usually follows the first batch of test calls. Most groups discover small location-specific exceptions here, and the value of a centralized system is that a rule fixed once applies consistently across every clinic that needs it.

Phase 4: Pilot on a subset of traffic

Before the AI owns the phones, it runs a controlled pilot. A common pattern is to start with one location or with overflow and after-hours calls, the volume that previously went to voicemail or rang out. This is low risk by design: any call the AI would have caught was a call no one was answering before.

During the pilot the team listens to recordings and transcripts, checks that appointments land correctly in the PMS, and confirms qualification details are captured. Conversion is watched closely, because the gap between a top performer and a lagging site, roughly 9 percent against under 3 percent in the BSD data, often traces back to scheduling rules and qualification handling that the pilot is meant to surface and correct.

The point of a pilot is not to prove the technology works in the abstract. It is to prove it works on your phones, with your providers, on your schedule, before it carries the full load.

Phase 5: Full go-live across locations

Once the pilot meets the agreed bar for booking accuracy and conversion, the AI takes the full inbound load. Many groups stage the rollout, bringing locations live in sequence rather than all at once, so each cutover can be verified and any last adjustments made before the next clinic goes live. The defining behavior change at go-live is coverage: every inbound call is answered, around the clock, with no hold and no voicemail.

The May 2026 figures show why staged proof matters. Across the five clinics the system logged roughly 66.5 hours of AI talk time and managed 350 total appointments in the first full month. That volume only converts to outcomes if the routing, training, and scheduling rules established in earlier phases hold up under real load.

Phase 6: Optimization as an ongoing loop

Go-live is the start of tuning, not the end of the project. With every location reporting into one command center, ownership can compare conversion, call volume, and appointments created side by side and act on the differences. A lower-converting site can be diagnosed, its rules adjusted, and the change measured against its peers. The estimated revenue impact, on the order of $129K in a month and roughly $1.5M annualized at an assumed $1,000 first-year value per new patient, is an illustration rather than a guarantee, but it underscores why the optimization loop is treated as a permanent part of the deployment.

One month is rarely enough to settle an argument about technology. It is enough to surface a pattern. May 2026 was the first full reporting cycle with the TensorLinks AI front desk live across all five Best Smiles Dental locations, and the numbers it produced are concrete enough to plan against. This section lays out the complete results: what the system handled, what it created, and what those figures mean for a five-location group trying to grow without adding headcount at every desk.

The Headline Totals

Across the five clinics in May 2026, the AI receptionist fielded 2,366 inbound calls and spent 3,993 minutes on the phone, roughly 66.5 hours of live conversation. From that activity it created 163 appointments and managed 350 appointments in total, a category that includes new bookings plus reschedules, confirmations, and routing of existing visits. 129 new patients were onboarded. The blended call-to-booking conversion rate came in at 6.9%.

Read those numbers in sequence and the operating story becomes clear. Every one of the 2,366 calls was answered. None rolled to voicemail, none sat on hold, and none arrived after hours to an empty office. Industry estimates commonly suggest that a busy dental practice misses a meaningful share of inbound calls during peak hours and loses most after-hours callers entirely. Whatever that figure was for Best Smiles before launch, in May it was effectively zero unanswered.

Per-Location Breakdown

The aggregate hides the variation that matters most to an operator. Here is the full distribution across the five sites for May 2026.

SouthSide carried the group. It generated more than a third of total call volume and, unusually, paired that volume with a strong 8.9% conversion rate. High volume and high conversion rarely travel together, because busier phones tend to mean more rushed handling. The AI does not get rushed, which is part of why the busiest site stayed efficient. Dr. C Advanced Dental posted the single best conversion at 9.1% on more modest volume, a sign that the call mix there skews toward ready-to-book intent. At the other end, NorthSide converted just 2.9% and StaplesMill 3.8%, and those two sites are where the near-term upside concentrates.

What Each Metric Actually Tells You

It helps to separate the operational metrics from the commercial ones.

• Calls (2,366) measure demand and coverage. This is the raw inbound load the front desk would otherwise absorb. Spread across roughly 21 business days, it averages well over 100 calls per day for the group, with SouthSide alone near 40 daily. That is a staffing burden the AI removed from the human teams.
• AI minutes (3,993) measure work performed. At about 66.5 hours of talk time, the system did the equivalent of more than a full-time week and a half of continuous conversation in a month, without breaks, sick days, or after-hours gaps. Average handle time lands near 1.7 minutes per call, which is brisk for a mix that includes scheduling, qualification, and routing.
• Appointments created (163) measure direct booking output. This is net-new schedule built by the AI.
• Appointments managed (350) measure total schedule touchpoints. The gap between 350 managed and 163 created, 187 appointments, reflects reschedules, confirmations, and routing the AI handled on existing visits. More than half of all appointment activity was schedule maintenance, the unglamorous work that quietly protects production by reducing no-shows and gaps.
• New patients (129) measure growth. These are the patients the practice did not have before, qualified and onboarded through the AI. This is the figure ownership cares about most, and it feeds the revenue model discussed later in this study.
• Conversion (6.9% blended) measures efficiency. It is the bridge between call volume and bookings, and it is the lever with the most room to move.

The most expensive call a practice ever takes is the one it never answers. Volume only becomes revenue once someone, or something, reliably picks up.

From Activity to Outcome

The commercial translation is deliberately conservative and should be read as an illustration, not a guarantee. Best Smiles values each new patient at roughly $1,000 in first-year production, a figure the client selected that sits within commonly cited ranges for general dentistry. Applied to 129 new patients, May produced an estimated ~$129,000 in first-year production value, which annualizes to roughly $1.5M across the five clinics if the month is representative. We treat that as an estimate throughout this study, because real production depends on case mix, acceptance, and retention that one month cannot confirm.

What May does confirm is the mechanism. Demand existed in the form of 2,366 calls. Coverage was complete. And a measurable share of that demand, 129 new patients and 163 created appointments, converted into schedule and growth. The spread between the best and weakest converting locations, 9.1% versus 2.9%, is the clearest signal in the data: the system is already producing, and the gap between sites is a roadmap rather than a verdict. The sections that follow examine those high and low performers in detail and quantify what closing that conversion gap could mean.

Across the five Best Smiles locations, two sites set the standard for what a well-run AI front desk produces. Best Smiles SouthSide proved the volume case: it absorbed the heaviest call load in the group and still converted at a high rate. Dr. C Advanced Dental proved the efficiency case: it posted the best conversion in the portfolio on a fraction of SouthSide's traffic. Together they form a two-part definition of "good" that the rest of the group can be measured against.

Best Smiles SouthSide: volume capture done right

SouthSide handled 809 inbound calls in May 2026, more than any other location and roughly a third of the group's total 2,366 calls. The AI receptionist logged 1,397 minutes of talk time at this site alone, again the most in the group. High volume is usually where front desks break down. Calls stack up, lines ring out, and the overflow lands in voicemail that rarely converts. SouthSide did not pay that tax. Every call was answered, and the location still produced an 8.9% call-to-booking conversion, well above the 6.9% blended average and second only to Dr. C.

The output followed. SouthSide created 72 appointments and managed 153 total — the largest managed book in the group by a wide margin. It onboarded 55 new patients, which is 43% of the group's 129 new patients for the month from a single site. The combination matters more than either figure alone. Plenty of locations can either run high volume or run high conversion. SouthSide did both at the same time, which is the harder result and the one most directly tied to growth.

What makes SouthSide instructive is that its strength is reproducible. The site is not converting well because of some quality the AI cannot extend elsewhere. It is converting well at scale because the front desk never becomes a bottleneck. When a high-demand location stops losing callers to hold times and after-hours gaps, the underlying demand simply shows up in the booked schedule. That is the volume-capture template: remove the capacity ceiling, and a busy phone turns into a busy chair.

Dr. C Advanced Dental: best-in-group conversion

Dr. C Advanced Dental tells the other half of the story. It took 395 calls, less than half of SouthSide's load, and used 667 minutes of AI talk time. On that smaller base it posted the highest conversion in the group at 9.1%, narrowly ahead of SouthSide and more than three times the rate of the group's weakest site. From those calls it created 36 appointments, managed 72, and onboarded 25 new patients.

The efficiency here is worth sitting with. Dr. C created 36 appointments from 395 calls. SouthSide created 72 from 809. The two sites convert callers into booked appointments at almost the same effectiveness per call, even though SouthSide is operating at roughly double the scale. Dr. C also onboarded 25 new patients, nearly half of its 72 managed appointments, which suggests a healthy mix of genuinely new demand rather than a book dominated by existing-patient churn.

The strongest front desks are not the ones that field the most calls. They are the ones where a higher share of every call that comes in ends as a confirmed appointment.

What we think drives the gap (stated as hypotheses)

The data tells us that SouthSide and Dr. C convert well. It does not, on its own, tell us why. The following are working hypotheses, framed as such, and worth validating before the group treats any of them as settled.

• Demand quality. Dr. C's 9.1% conversion on lower volume may reflect a caller base with clearer intent — more callers who already mean to book rather than to ask a general question. If true, the lever is less about the AI and more about which marketing and referral sources feed each line.
• Schedule availability. Conversion is capped by what the AI can actually offer. A site with open slots and flexible provider hours converts more of the same calls than a site with a thin book. Both SouthSide and Dr. C likely had enough bookable supply to say yes when a caller was ready.
• Clean scheduling logic at the location. The AI books against each clinic's own calendar and rules. Where appointment types, provider mapping, and new-patient slots are well configured, qualification and booking complete in one call. Configuration quality is a plausible differentiator between the high performers and the lagging sites.
• Volume that exposes capacity, not quality. SouthSide shows that high call counts do not depress conversion when no calls are dropped. That argues the binding constraint at busy sites is capacity, not the front desk's ability to convert.

Why these two become the group template

SouthSide and Dr. C give ownership two reference points instead of one. SouthSide is the answer to "what happens when our busiest phone never goes unanswered" — 55 new patients, the deepest managed book in the group, conversion held above average at peak load. Dr. C is the answer to "how high can conversion go when the inputs line up" — 9.1%, the per-call benchmark every other location can be held to.

For a five-location group, the practical move is to treat the 8.9–9.1% conversion band as the internal target and work backward at the sites that fall short. NorthSide converted at 2.9% and StaplesMill at 3.8% in the same month on the same platform. The gap between those sites and the high performers is not a software gap, since the AI is common across all five. It points to location-level factors — demand mix, schedule availability, configuration — which are the same hypotheses above and the same things a group can actually change. The high performers do more than carry the month's numbers. They define a concrete bar and narrow the search for what is holding the rest of the portfolio back.

The five clinics did not move in lockstep during May 2026. Two pulled ahead. Three sat in the middle or back of the pack, and those three are where the most attainable upside lives. GlenAllen, StaplesMill, and NorthSide together fielded 1,162 inbound calls in the month and converted 67 new patients out of that volume. They were not failing. They were leaving bookings on the table that the same software, tuned the same way, was already capturing elsewhere in the group.

The useful comparison is internal. SouthSide converted calls to bookings at 8.9 percent and Dr. C Advanced Dental at 9.1 percent. The group blended out to 6.9 percent. When a clinic in the same portfolio, on the same AI front desk, runs at less than half the conversion of its sibling locations, the gap is rarely about patient demand. The phones are ringing. The question is what happens after the connection.

GlenAllen: a near-leader hiding in the middle

GlenAllen is the strongest of the three and arguably the most encouraging. On 338 calls it created 26 appointments and onboarded 20 new patients at a 7.7 percent conversion rate. That sits just below the two leaders and well above the blended group average. The shape of GlenAllen's month suggests the call handling and scheduling flow are largely working; this is a fine-tuning case, not a turnaround case.

The illustrative headroom is modest but real. If GlenAllen reached the leader band of roughly 9 percent on its existing 338 calls, expected appointments created would rise to about 30, an estimated gain of four. At the client's chosen value of about $1,000 in first-year production per new patient, that is on the order of $4,000 in additional monthly value, or roughly $48,000 annualized. Treat both figures as estimates. The practical lever here is likely small: tightening which call types convert, confirming the schedule has open new-patient slots when the AI tries to book, and making sure follow-up fires on the handful of callers who hesitate.

StaplesMill: high volume, low yield

StaplesMill is the most important of the three, because it has the second-highest call volume in the entire group and one of the lowest conversion rates. On 548 calls it created just 21 appointments and onboarded 21 new patients, a 3.8 percent conversion. The demand is plainly there. More people called StaplesMill than called Dr. C Advanced Dental and GlenAllen combined, yet StaplesMill produced fewer created appointments than either.

This is the classic profile of a conversion problem rather than a marketing problem. When volume is high and yield is low, the friction usually sits in one of a few places: callers who want something the booking flow cannot smoothly resolve, schedule availability that does not match when the AI tries to place a patient, or a follow-up gap on callers who do not book on the first contact. The data alone cannot prove which of these dominates at StaplesMill. What it can do is size the prize.

These are estimates derived only from StaplesMill's own 548 calls multiplied by each target rate. Even the most conservative scenario, simply pulling StaplesMill up to the group's own blended average, points to roughly 17 additional appointments a month from the volume the clinic is already receiving. Applied to the client's $1,000 per-new-patient framing, the blended-average scenario implies somewhere near $17,000 in additional estimated monthly value, and the leader-band scenario closer to $28,000. The numbers move fast at StaplesMill precisely because the call volume is so high; a few points of conversion compound against a large base.

NorthSide: smallest base, biggest percentage gap

NorthSide had the group's lowest call volume and its lowest conversion. On 276 calls it created 8 appointments and onboarded 8 new patients at 2.9 percent. In absolute terms the shortfall is smaller than StaplesMill's. In relative terms it is the widest gap in the portfolio, running at roughly a third of the group's blended rate and less than a third of the leaders'.

Low volume plus low conversion is a different diagnosis than StaplesMill. Here the priority order may flip: with fewer calls coming in, every connected call matters more, so call handling quality and same-call booking carry outsized weight, and there may also be a top-of-funnel question about why volume is low to begin with. On the conversion side, the illustrative math is still meaningful. At the group blended 6.9 percent, NorthSide's 276 calls would be expected to produce about 19 created appointments, an estimated gain of 11. At the leader band of roughly 9 percent, that rises to about 25, a gain of 17. In the client's revenue framing those are estimated monthly gains of roughly $11,000 and $17,000 respectively, again labeled as illustration rather than booked revenue.

When two clinics on identical software convert at twice the rate of a third, the variable is rarely the market. It is what happens in the seconds after the phone connects.

What the three sites share

Across GlenAllen, StaplesMill, and NorthSide the common thread is that the inbound demand is being answered but not fully converted. The centralized AI front desk has already removed the most expensive failure mode, the unanswered or voicemail-bound call, so the remaining gap is downstream: matching availability, qualifying new patients cleanly, and following up with callers who do not commit on first contact. Summed conservatively at the group's own blended rate, lifting these three clinics off their May baselines points to well over two dozen additional created appointments per month, all from calls the group is already receiving. Every figure above is an estimate built from the May 2026 data and the client's $1,000 per-patient assumption, and should be read as directional rather than guaranteed.

Every operating expense in a dental group eventually faces the same question from ownership: what did it return? For a centralized AI front desk, the answer starts with a single number from May 2026. Across the five Best Smiles Dental locations, the system onboarded 129 new patients in one month. That figure is the foundation of the revenue model, and unlike softer benefits such as shorter hold times or cleaner schedules, it converts to dollars in a way owners and DSO finance teams can audit.

Why new patients are the right unit of measure

The 129 new patients are the cleanest line to monetize because new-patient acquisition is where the front desk most directly touches revenue. A missed new-patient call is not a deferred appointment that reschedules itself. It is usually a permanent loss to a competitor down the road. Rescheduling, routing, and after-hours coverage all contribute value too, but they are harder to price without double-counting. Anchoring the model on new patients keeps the math conservative and defensible.

To translate patients into revenue, you need a per-patient value. Best Smiles Dental chose to value each new patient at approximately $1,000 in first-year production. This is an estimate, and we treat it as one throughout. It reflects a typical mix of exams, imaging, hygiene, and the early restorative or treatment-plan work a new patient tends to accept in their first twelve months. Industry estimates for first-year new-patient value commonly land in this neighborhood, though the true figure varies by payer mix, fee schedule, case-acceptance rates, and the share of patients who proceed with larger treatment.

The base case

At roughly $1,000 per new patient, 129 new patients in May 2026 represent an estimated $129,000 in first-year production for that month's cohort. Held flat across twelve months, that points to an estimated $1.5 million in annualized new-patient production attributable to patients the AI front desk onboarded. Both numbers are illustrative estimates, not booked revenue, and they assume May is broadly representative of a normal month rather than a seasonal peak or trough.

A few caveats keep the base case honest. First-year production is not collected revenue; it is the value of work performed, before adjustments and write-offs. The annualized figure assumes the system continues onboarding new patients at a similar pace, which depends on call volume, marketing spend, and capacity at each clinic staying roughly constant. And the model attributes the new patients to the AI front desk because it is the channel that answered the calls and booked the appointments. It does not claim the AI generated the demand. Marketing, reputation, and referrals create the inbound call; the front desk's job is to convert it rather than lose it.

Sensitivity: three values per new patient

Because the per-patient value is the single most important assumption, it deserves a stress test. The table below holds the 129 new patients fixed and varies only the estimated value per patient: a conservative $600, the $1,000 base case, and an aggressive $2,500 that approximates multi-year lifetime value rather than first-year production. Every figure is an estimate for illustration.

The spread matters. Even the conservative scenario, which assumes a thin first-year value and ignores any work beyond the first twelve months, points to an estimated $77K per month and nearly $929K annualized. The aggressive column is not a forecast. It simply shows what the same 129 patients are worth if you account for the years of hygiene recalls and follow-on treatment a retained patient typically delivers. Most groups will plan against the conservative-to-base range and treat anything above it as upside.

The decisive economic event at a dental front desk is not the appointment that gets booked. It is the new-patient call that never gets answered, because that loss compounds for years.

Payback and why this is a high-ROI line item

The reason a centralized AI front desk reads as a high-ROI investment is the relationship between its cost and the value it protects. We are not publishing Best Smiles Dental's specific contract terms, but the structure of the argument holds across pricing models. A front-desk solution is a fixed or usage-based monthly cost. The new-patient production it captures is recurring and, in the base case, lands around $129K per month in estimated first-year value. For the line item to fail to pay for itself, the monthly cost would have to consume a very large share of that estimated value, which is far outside how these systems are typically priced.

Put differently, payback is best framed in patients, not just dollars. If the service costs the equivalent of a handful of new patients per month, then once the system onboards beyond that handful, every additional new patient is contribution toward production rather than cost recovery. At 129 new patients in a single month across five locations, that breakeven point is reached early and cleared by a wide margin. The same logic explains why the lowest-converting locations still carry a strong case: even NorthSide's 8 new patients and StaplesMill's 21, valued conservatively, comfortably exceed a per-location cost typical of this category.

There is also an avoided-cost angle that the table does not capture. The 129 new patients were onboarded against roughly 66.5 hours of AI talk time spread across 2,366 inbound calls, including nights and weekends, with no hold queue and no voicemail. Matching that coverage with additional human staff across five front desks would carry its own salary, benefits, and management overhead. The revenue model above is deliberately built on captured production alone, so any labor savings sit on top of it as further upside rather than being folded into the headline estimate.

Reading the model honestly

To summarize the assumptions in one place: the 129 new patients are actual May 2026 results; the per-patient values are estimates; the monthly figures assume those values are accurate; and the annualized figures assume twelve comparable months. The dollar outputs are illustrative production estimates, not collected revenue or guaranteed forward results. Even with every one of those caveats applied, and even in the conservative column, the model lands in a range where the captured new-patient value is several multiples of a typical front-desk cost. That is the core of the ROI case: not a single optimistic headline number, but a result that stays compelling across a wide band of assumptions.

Across the five locations, conversion ranged from 9.1% at Dr. C Advanced Dental down to 2.9% at NorthSide. That spread is the most actionable number in the entire engagement. SouthSide and Dr. C are not staffed by fundamentally different people or serving fundamentally different patients. They are running a better version of the same job. The opportunity is not to discover a new growth lever. It is to make every clinic perform the way the best ones already do.

A centralized AI front desk turns that aspiration into a system. When the same answering logic, the same qualification script, and the same routing rules run behind every phone number, the gap between top and bottom stops being a function of who happens to be at the desk on a Tuesday afternoon. It becomes a configuration you can tune once and apply everywhere.

What separates a 9% location from a 3% location

The leaders convert because they execute four things consistently. The laggards lose ground because those same four things break down under pressure, after hours, or during the front desk's busiest stretches.

• Instant answer. Every call is picked up on the first ring with zero hold and no voicemail. A caller who reaches a person, or an AI receptionist indistinguishable in responsiveness, never gets the chance to dial the next practice on their list.
• Consistent call handling. The same greeting, the same questions, and the same path to a booked slot on call number one and call number four hundred. Quality does not decay as volume climbs.
• Structured new-patient capture. Insurance, reason for visit, and preferred timing collected the same way every time, so a high-intent caller is qualified and scheduled rather than told someone will call back.
• No after-hours drop-off. Calls that land at 7 p.m. or on a Saturday get booked instead of abandoned. For many groups, evenings and weekends are where the lagging numbers actually come from.

NorthSide handled 276 calls and converted 8, while Dr. C handled 395 calls and converted 36. The volume difference is real, but it does not explain a 3x conversion gap. What explains it is consistency of execution, and consistency is precisely what a centralized system is built to deliver.

The repeatable playbook

Lifting a trailing location is a sequence, not a single switch. The work breaks into clear steps that the command center makes visible and repeatable.

1. Baseline against the leaders. Compare each location's conversion, answer rate, and after-hours capture to SouthSide and Dr. C. The leaders become the internal benchmark, not an external estimate.
2. Clone the winning configuration. Take the qualification flow and routing logic running at the highest-converting site and apply it as the default everywhere. There is no reason every clinic should reinvent the script.
3. Close the after-hours gap. Confirm that 24/7 answering is live and that overnight and weekend calls flow into the same booking path as daytime calls.
4. Review the misses. Use call records to find where conversations stall, then refine the prompts and handoff rules. Each fix propagates to all five locations at once.
5. Hold the standard. Track weekly. When a location drifts, the data surfaces it before a month of bookings is lost.

The fastest way to grow a multi-location group is rarely a new location. It is closing the distance between your best front desk and your average one.

The illustrative upside

Consider what convergence toward the leaders could mean. These figures are estimates for illustration, built on May 2026 call volumes and the client's chosen $1,000 first-year new-patient value. Actual results will vary with demand, capacity, and case mix.

If NorthSide and StaplesMill alone moved from their current rates to roughly the group's 6.9% blended level, that is on the order of 28 additional bookings a month from existing call volume, with no new marketing spend. Carry the new-patient share through at the client's illustrative $1,000 value and the monthly upside lands in the mid five figures, again as an estimate rather than a guarantee. Push the trailing sites toward Dr. C's 9.1% and the figure climbs further.

The point is not the exact dollar amount. It is that the calls are already arriving. The leaders prove the demand converts when the system around it is consistent. A centralized AI front desk is the mechanism that copies that consistency from the best location to the rest, then keeps it from slipping back. Closing the conversion gap is less about working harder at the desk and more about removing the variability that lets a 9% site and a 3% site coexist inside the same group.

Capacity is the quiet variable behind every front-desk metric. A practice can have the friendliest team and the best schedule logic, but if no one picks up, the call is gone. In May 2026, the TensorLinks AI front desk fielded 2,366 inbound calls across Best Smiles Dental's five locations and produced 3,993 minutes of talk time, roughly 66.5 hours. The number that matters most is not the total minutes. It is the number of calls that hit a busy signal, a hold queue, or voicemail: zero.

Sixty-six hours of talk time, none of it queued

Spread across one month and five clinics, 66.5 hours of conversation does not sound dramatic. The challenge is that call volume is not evenly distributed. Inbound dental calls cluster: the first hour after open, the lunch window when patients step away from their own jobs, and the late afternoon when families coordinate after school. During those peaks, a human front desk with one or two people physically cannot answer two or three lines at once while a patient is standing at the counter and another is mid-checkout.

An AI receptionist has no single-threaded constraint. The system answered the 809 calls into SouthSide and the 548 into StaplesMill without either location ever competing for the same set of hands. Every concurrent caller got a live answer. There was no "please hold," no rollover to a shared voicemail box, no prospective patient deciding, in the eight seconds it takes to reach a recording, to dial the practice down the street instead.

What zero voicemail is worth

The economics of voicemail are unforgiving, and they are worth stating plainly because they are usually invisible on a P&L. Industry estimates commonly suggest that a meaningful share of inbound calls to busy practices go unanswered during business hours, and that a large majority of callers who reach voicemail simply hang up rather than leave a message. New patients are the least patient of all. Someone shopping for a dentist, often in mild pain or with a specific scheduling need, treats a voicemail prompt as a closed door.

That is the gap TensorLinks closes structurally rather than through effort. Of the 2,366 calls handled, BSD onboarded 129 new patients and created 163 appointments. Those new patients did not come from the calls a team answered on a good day. They came from answering all of them, including the ones that historically would have been lost to a full queue or an after-hours recording.

A missed call is not a deferred opportunity. For a prospective patient, it is usually a decision to call someone else, and you rarely get a second chance to make it.

Evenings and weekends stop leaking

The second capacity gain is temporal. A traditional front desk is open maybe 40 to 45 hours a week. The rest of the week, including evenings, weekends, and holidays, the phone either rings out or lands in voicemail. Yet that is precisely when many working adults have time to call about their own care or their family's.

Because the AI answers 24/7, those off-hours calls became bookable conversations instead of lost ones. A caller at 8:40 p.m. could be qualified as a new patient and placed on the schedule on the spot, with the appointment flowing into that clinic's own calendar. No one had to return the call the next morning, by which point a meaningful fraction of those callers have already booked elsewhere. The practice did not extend its staffed hours by a single minute, but its reachable hours became continuous.

Tying capacity to captured patients

It helps to connect the talk-time figure to outcomes rather than treating it as a usage stat. The table below frames the chain from raw capacity to result.

At the client's chosen estimate of roughly $1,000 in first-year production per new patient, those 129 onboardings represent on the order of $129K for the month. That figure is an illustration, not a billed amount, and it depends on the practice's case mix and follow-through. The directional point holds regardless of the exact multiplier: a large portion of that value existed only because every call was answered, including the ones a staffed desk could never have reached.

Capacity as a fixed asset, not a staffing scramble

For a multi-location group, the strategic shift is that capacity stops being something you buy in headcount increments and start managing as a constant. Adding a sixth or seventh location does not require re-solving the "who covers the phones at 6 p.m." problem each time. The same AI front desk absorbs the new volume, peaks included, with no degradation in answer rate. For ownership, that turns front-desk capacity from a recurring operational risk into infrastructure that simply scales with the portfolio.

The financial case for a centralized AI front desk is the headline, but it is not the whole story. For an owner running five locations, the day-to-day operational reality is what determines whether a system gets adopted or quietly abandoned. At Best Smiles Dental, the clearest operational shift in May 2026 was where staff attention went. The AI handled 2,366 inbound calls and roughly 66.5 hours of live talk time across the group. That is time front-desk teams did not spend tethered to a ringing phone, and it changed the texture of the workday at the desk.

Reallocating staff time toward the patient in the chair

A dental front desk runs two jobs at once. One faces the phone. The other faces the patient standing at the counter trying to check in, settle a balance, or ask about a treatment plan. Those two jobs compete, and the phone usually wins because a ringing line feels urgent. The result is a familiar scene: a patient waiting while the coordinator finishes a scheduling call, or a call sent to voicemail because the lobby got busy.

When the AI answers every inbound call, that competition eases. Across BSD's five locations, the system fielded the full call volume without putting anyone on hold and without routing to voicemail. The front-desk team was no longer the single point of failure for inbound contact. Routine work — booking a cleaning, confirming an appointment, answering hours and location questions, qualifying a new patient — moved to the AI, while staff focused on the higher-value, in-person interactions that genuinely need a human: the nervous patient, the complex financial conversation, the treatment coordination that closes a case.

This is reallocation, not replacement. The people stay. What changes is what they spend their attention on. For a busy site like SouthSide, which alone took 809 calls and produced 1,397 minutes of AI talk time, absorbing that load at the desk would have meant either constant interruption or a dedicated phone person. Neither serves the patient standing in the lobby.

Reducing front-desk strain and turnover pressure

Front-desk burnout is a real operating cost in dental groups, and it rarely shows up as a clean line item. It surfaces as turnover, training time for the next hire, and the dip in service quality while a new person learns the practice. Phones are a large part of that strain. The interruptions are relentless, the work is repetitive, and missed calls create downstream cleanup that lands back on the same person.

By taking the repetitive, high-interruption portion of the job off the team's plate, a centralized AI front desk lowers the ambient pressure of the role. The coordinator who is not bracing for the next ring tends to be more present with the patient in front of them. Volume spikes — a Monday morning, a post-holiday rush, a single location getting slammed — no longer translate directly into more stress at the desk, because the AI scales to the call volume without anyone scrambling to keep up.

The most overlooked benefit of automating the phone is not the calls it answers. It is the attention it gives back to the people at the desk.

One consistent patient experience across five locations

Multi-location groups have a quiet consistency problem. Each front desk develops its own habits, its own phone manner, its own approach to handling a new-patient inquiry. A caller's experience can depend on which location they happened to dial and who happened to pick up. For a brand trying to feel like one practice rather than five, that variance is a liability.

A centralized AI front desk standardizes the first impression. Every caller, at every location, reaches the same calm, informed point of contact. New-patient qualification follows the same script and captures the same information whether the call lands at NorthSide or Dr. C Advanced Dental. Scheduling logic, routing rules, and the tone of the interaction are uniform by design, while each clinic keeps its own schedule and local routines. The patient gets a consistent experience; the location keeps its operational independence.

That consistency is also measurable in availability. Because the AI answers around the clock, the experience does not degrade after hours, during lunch, or when a location is short-staffed. A caller at 9 p.m. gets the same instant answer as one at 9 a.m.

What the patient actually feels: instant, available, on their schedule

From the patient's side, the improvements are simple and immediate. Three things change:

• No hold, no voicemail. Every call is answered live. Industry estimates commonly suggest a meaningful share of inbound calls to dental practices go unanswered during peak hours, and a voicemail rarely converts into a booked appointment. Removing both removes the most common reason a prospective patient gives up and calls the next practice on the list.
• Booking on their own schedule. Patients can schedule by voice, online, or text, at whatever hour suits them. Many people simply cannot call during business hours, and that constraint quietly costs practices appointments. Unified scheduling across channels meets patients where they are.
• Speed and certainty. A new patient can be qualified, slotted, and confirmed in a single interaction, rather than playing phone tag across a day or two. That immediacy matters most for the new patients who drive growth — 129 of them onboarded across the group in May 2026.

None of this requires inventing a glowing quote to make the point. The operational logic stands on its own: when the phone is always answered, when the desk is freed to focus on the lobby, and when the experience is identical across every location, the practice runs calmer and the patient experience gets better at the same time. The revenue follows from that, but the operational and human gains are worth counting in their own right.

Run a five-location dental group with five independent front desks and you face a structural problem: every clinic produces its own version of the truth. One office tracks new patients in a spreadsheet, another counts them differently, a third does not count them at all. By the time numbers reach ownership, they are weeks old and shaped by whoever assembled them. You can feel a location underperforming long before you can prove it, and by then the lost bookings are gone for good.

The TensorLinks portfolio command center exists to close that gap. Every inbound call, every booking, every new-patient qualification flows through the same AI front desk, which means it is also recorded the same way. Ownership gets one screen showing calls, conversion, appointments created, appointments managed, and new patients onboarded across all five clinics, updated as the month unfolds rather than reconstructed after it ends. The May 2026 figures in this case study are a direct read from that view, not a manual compilation. That distinction matters: when the data collects itself, comparison becomes trustworthy.

Why one view changes the conversation

The argument for portfolio visibility is written into Best Smiles Dental's own numbers. Across the five locations, call-to-booking conversion ranged from 2.9% at NorthSide to 9.1% at Dr. C Advanced Dental. That is more than a 3x spread inside a single brand, operating under one owner, in one market. Without a common system, a spread like that hides in plain sight. Each location looks busy. Each front desk feels productive. Only when the conversions sit side by side in the same units does the pattern surface.

Consider what the command center made legible in month one:

• Volume and conversion are not the same lever. SouthSide handled the most calls (809) and still converted at 8.9%, while NorthSide handled 276 calls and converted at 2.9%. High volume did not guarantee results, and lower volume did not excuse weak ones.
• The strongest clinic was not the biggest. Dr. C Advanced Dental led on conversion at 9.1% from 395 calls. In a federated reporting world, a smaller office quietly outperforming the flagship is exactly the signal that never reaches the owner.
• The weak points are specific, not vague. StaplesMill fielded 548 calls but converted only 3.8%, the kind of mismatch that points to a fixable issue rather than a busy month.

None of these observations required a consultant or a quarterly review. They came from one screen, in one set of units, in close to real time. That is the practical meaning of enterprise visibility: the owner can act on variance while the month is still in progress instead of explaining it after the fact.

Local routines stay local

Visibility alone is not the goal. Plenty of multi-location operators have tried to centralize their way to consistency and ended up flattening the things that made each office work. A practice that has spent years training its team on a particular recall cadence, a specific way of handling new-patient questions, or a schedule built around its providers does not want that erased by a head-office standard.

The command center is deliberately built so it does not have to be. Each clinic keeps its own schedule, its own provider templates, its own booking rules and routines. The AI front desk answers every call, qualifies new patients, and routes or reschedules according to that location's setup, not a one-size-fits-all script imposed from above. SouthSide's calendar logic and NorthSide's calendar logic remain distinct. What is shared is the standard of capture and the standard of measurement, not the local workflow.

This is the dual capability that multi-location groups and DSOs actually need. Centralized standards give ownership comparability, accountability, and a single source of truth. Local autonomy keeps each practice running the way it knows how to run. The two are usually treated as a trade-off. Pick centralization and you lose the local nuance; pick autonomy and you lose the portfolio view. The front desk is the rare layer where you can hold both, because the same engine can enforce a consistent way of answering and recording calls while still respecting five different schedules underneath.

The hardest problem in a multi-location group is not running five clinics. It is seeing them as one portfolio without forcing them to behave as one clinic.

What this unlocks for the operator

Once variance is visible and locations keep their autonomy, management shifts from anecdote to allocation. An owner can look at the spread between Dr. C Advanced Dental and NorthSide and ask a targeted question: what is the high performer doing at the point of the call that the low performer is not, and can that be propagated without disrupting either office's schedule? Coaching gets aimed where the return is largest. Marketing spend can be weighed against the conversion that follows it, not just the calls it generates.

For a group eyeing its next acquisition, the value compounds. A new location plugs into the same command center on day one and starts reporting in the same units as the existing five, with no integration of disparate front-desk habits required. The portfolio scales without the reporting chaos that usually scales with it. That is what "enterprise visibility meets local efficiency" means in operation: ownership sees the whole, each clinic keeps its own, and growth stops being a step backward in clarity.

Best Smiles Dental's first month with TensorLinks AI offers more than a single group's results. It surfaces patterns that apply to almost any dental organization running multiple front desks. The numbers from May 2026 (2,366 calls answered, 163 appointments created, 129 new patients onboarded, and a 6.9% blended call-to-booking conversion) are useful, but the lessons underneath them travel further than the specifics.

Six lessons for multi-location dental groups

1. Your conversion problem is mostly an answer-rate problem. Before you redesign scripts or retrain staff, count the calls nobody picks up. BSD's locations did not need more persuasive receptionists as much as they needed someone available for every ring. When the same AI handled inbound calls across all five clinics, conversion ranged from 2.9% to 9.1%. That spread is the real opportunity, and it only becomes visible once every call is actually answered and logged.
2. Variation across locations is the asset, not the embarrassment. A 9.1% conversion clinic and a 2.9% conversion clinic running on the same platform is a built-in benchmark. The high performers show what the qualification and booking flow can do when local conditions cooperate. Treat the gap as a roadmap for coaching and scheduling adjustments rather than a reason to apologize.
3. Volume and conversion are separate dials. The busiest location was also a strong converter, while the second-busiest converted at less than half that rate. High call counts do not guarantee bookings, and modest call counts can still produce excellent yield. Measure both per location, because a group-wide blended average hides the clinics that need attention most.
4. Centralize the answer, keep the schedule local. The result that made the model work was structural. One AI front desk fielded every call, but each clinic retained its own calendar, providers, and routines. Owners gained a portfolio view without forcing five practices into one rigid workflow. For groups that grew by acquisition, this preserves the local character patients already trust.
5. After-hours and overflow capacity is found revenue. Roughly 66.5 hours of AI talk time in a month included calls that previously would have hit voicemail or a busy signal. Industry estimates commonly suggest a large share of missed dental calls never convert, since callers simply dial the next practice. Capturing even part of that demand changes the math without adding a single new patient source.
6. Decide your value-per-patient framing before you report ROI. BSD chose to value each new patient at roughly $1,000 in first-year production. That assumption, applied to 129 new patients, produces an estimated $129K per month and about $1.5M annualized. The figure is an illustration, not a guarantee, and your own number will depend on your case mix and treatment acceptance. Pick the assumption deliberately and label it as an estimate every time you cite it.

Frequently asked questions

How accurate is an AI receptionist on real calls?

Accuracy is best judged by outcomes, not demos. Across BSD's five locations the AI created 163 appointments and managed 350 in a single month, while qualifying new patients and routing reschedules. The system is built to confirm details, capture intent, and hand off cleanly when a call falls outside its scope. We recommend reviewing transcripts and booking accuracy during your first weeks rather than relying on a vendor's word.

Will patients accept talking to an AI instead of a person?

Most callers care more about being answered immediately than about who answers. The alternative to AI reception is frequently a voicemail box or a hold queue, both of which patients dislike far more. The AI answers 24/7 with zero hold time, books in the same conversation, and escalates when a human is genuinely needed. In practice the experience often feels faster than the old front desk, especially after hours.

How does this integrate with the systems we already run?

TensorLinks AI is designed to sit on top of your existing scheduling rather than replace it. Each location keeps its own calendar and routines while the AI reads availability and writes appointments back. Because the platform is centralized, you do not have to standardize all five practices onto one workflow to get started. Integration scope is confirmed during kickoff so each clinic's setup is reflected accurately.

What does it cost, and how do we know it pays off?

The honest answer is that ROI depends on your own value-per-patient assumption. Using BSD's illustrative $1,000 first-year figure, 129 new patients in one month points to roughly $129K in estimated production against the cost of a centralized AI front desk. Even with a conservative value per patient and a discount for patients who would have booked anyway, the captured after-hours and overflow demand tends to carry the case. Run the numbers with your own inputs before deciding.

How is patient data handled from a security and compliance standpoint?

Patient information in dental settings is protected health information, so any front-desk system must be built with that in mind. TensorLinks treats data handling, access controls, and call records as core requirements rather than afterthoughts. We are happy to walk your administrator and any compliance advisor through how information is processed and safeguarded at a level appropriate to your obligations. Treat this as a standard part of your evaluation.

How quickly will we see results?

BSD's headline figures came from the first full reporting month. Time-to-value depends on how fast each location's scheduling details are confirmed during onboarding, but the model is designed to start answering and booking calls early rather than after a long ramp. Expect the first weeks to also generate the data you need to spot conversion gaps between locations.

The path forward

Scaling a dental group has long meant scaling the front desk, and the front desk has long been the part that breaks first. A centralized AI receptionist changes that relationship. Adding a sixth or tenth location no longer requires standing up another phone team and hoping it answers consistently. The same answer layer extends across the portfolio while each new clinic keeps the local rhythm its patients already know.

The groups that win the next decade will be the ones that stop treating every missed call as a cost of doing business and start treating it as recoverable revenue.

That is the shift BSD made in a single month, and the figures here are an early, estimated snapshot rather than a ceiling. If you run more than one location and suspect your real conversion problem is hiding in the calls you never answer, the next step is to see the command-center view against your own numbers. Book a demo and we will map what a centralized AI front desk could look like across your locations, with revenue framing built on your assumptions rather than ours.