01. The Problem
Where the practice actually leaks.
A typical 3-chair general practice grosses $1.2M to $1.8M and thinks it is full. It is not full. It is leaking. The leak is invisible because the front desk is busy, the schedule looks tight, and the hygienists are booked two weeks out. None of that is the same as a healthy practice. The leak usually shows up in three places, every month.
New-patient calls land between 5pm and 9pm, on weekends, and during lunch. Voicemail is not a recovery channel. The caller dialed three other practices on Google in the next 90 seconds.
Industry recall sits between 35 and 45 percent. Patients fall off the 6-month and 4-month perio maintenance schedules. Postcards do not move the number. SMS with a real booking link does.
Illustrative math: one canceled crown is $1,400, one no-show hygiene block is $180. An assumed 14 percent no-show rate on a $1.5M practice puts about $210K of recoverable production at risk every year.
Why the front desk cannot fix this alone.
Most operators try to patch with a third front-desk hire. That seat costs $52K plus benefits, takes 90 days to ramp, never works evenings or weekends, burns out inside 14 months, and still cannot answer two lines at once. The math does not work and the labor pool is shrinking.
The other common patch is a $99 a month online scheduler bolted onto the PMS. It catches a sliver of self-scheduling new patients and does nothing for recall, no-show recovery, eligibility, or reviews. A single point solution against a five-point problem. The Nirvani stack treats the practice as a complete operating system. Humans stay focused on chairside coordination and treatment-plan presentation. The AI layer handles the round-the-clock administrative work.
02. After the Stack
What the dashboard looks like at day 90.
Illustrative targets, not measured results. These are modeled day-90 figures for a 3-chair practice, built from the assumptions on this page. Your timing depends on your PMS, your patient base, and your starting numbers.
Recall reactivation is the headline. Going from an assumed 38 percent baseline to the 78 percent target roughly doubles the hygiene book without adding a chair, a hygienist, or changing the production mix. Hygiene is the most profitable operatory and the diagnostic on-ramp for restorative, perio, ortho, and cosmetic.
03. The Stack
Five modules. A 90-day plan to fully live.
The stack ships in a specific order because each module unlocks data infrastructure the next one needs. We deploy in sequence, not parallel. Module 01 is always first because recall is the fastest path to booked production, and it is built to help pay for the rest of the deployment.
Each module below contains four things. What it does. The mechanism (voice agent call flow, SMS sequence, PMS integration points). The exact sample script. And an illustrative target with the numeric impact you can track in the daily huddle.
AI Hygiene Recall Engine
What it does. Pulls every patient from your PMS whose last prophylactic or perio maintenance visit was 6 months ago (standard recall) or 4 months ago (perio maintenance). Filters out future-booked, deceased, moved, and opted-out. Works the list across three channels until the patient books, replies STOP, or hits the dead-list threshold.
Why it matters. The hygiene book is the highest-margin chair and the diagnostic on-ramp for everything else. A 6-month hygiene patient is 3.4x more likely to accept a restorative treatment plan than a lapsed one. Moving from 38 percent toward the 78 percent target is the single highest-leverage move a practice can make.
Why postcards do not work. Sub-1 percent response. The office manager gets through to 23 percent of callbacks. The patient who answers is mid-day, distracted, and says "I'll call back." She doesn't. The Nirvani stack works the list while the office is closed, lets the patient tap a link, and presents three slots that fit her schedule.
The mechanism
- Day 0. Nightly sync pulls the recall list from your PMS (Dentrix, Eaglesoft, Open Dental, Curve, Denticon). Eligibility flagged against active insurance and last visit type.
- Day 1 AM. Outbound AI voice call to the primary number, 9am to 6pm patient local time. Agent introduces by practice name, confirms identity, offers three slots matching preferred chair time.
- Day 1 PM. If voice did not reach a human, SMS sequence triggers. Short, branded, with a one-tap booking link. A2P 10DLC registered with prior-relationship consent.
- Day 4. Follow-up SMS, different angle (insurance benefits, year-end reminder, hygienist availability).
- Day 10. Final outbound voice call, different time of day. After this, the patient sits in a 90-day cool-off list.
- Booking. Slot writes back to PMS via direct integration. Confirmation text fires immediately. Morning huddle dashboard updates in real time.
Sample voice script
Illustrative target
24/7 New-Patient Intake
What it does. Routes every new-patient call that arrives outside office hours, during lunch, or while both front-desk lines are busy, to the AI voice agent. The agent answers immediately, runs a structured intake (reason for visit, carrier and member ID, urgency, preferred chair time), books to the new-patient operatory, sends confirmation SMS, and posts a summary to the morning huddle dashboard.
Why it matters. 47 percent of dental new-patient calls happen outside core 9-to-5 office hours. Your two largest competitors are also in the Google Maps 3-pack and the prospective patient is dialing three of you in a row. Whoever answers first, books. The conversion gap between "answered live" and "left a voicemail and got called back next day" is roughly 8x.
Not the same as a virtual receptionist. A human virtual receptionist costs $1,200 to $2,800 per month, only handles scripted intake, cannot write back to your PMS in real time, hands off in a queue, and still gets benefits questions wrong half the time. The AI agent handles unlimited concurrent calls, writes directly to your appointment book, and delivers the recording and transcript to the morning huddle.
The mechanism
- Call routing. Front desk rings for 4 rings during office hours. If unanswered or after hours, the call rolls to the AI agent. No IVR maze. No "press 1 for new patients."
- Adaptive intake. Agent identifies new vs existing patient in the first 8 seconds. New gets the 11-field intake. Existing routes to appointment, billing, or message-the-doctor.
- Insurance capture. Agent collects carrier, member ID, and primary subscriber DOB. Stores to the new-patient record for eligibility (module 03).
- Booking write-back. Direct PMS integration writes to the correct operatory column with correct visit type and procedure codes pre-loaded (D0150 comprehensive exam, D0210 FMX, D1110 prophy as default new-patient block).
- Handoff exceptions. Emergencies, pediatric on adult-only practices, and "I want a human" requests immediately page the on-call team member with callback number and summary.
Sample voice script
Illustrative target
Insurance Eligibility Pre-Screening
What it does. Captures insurance during the booking call, runs eligibility against the carrier (Delta, Cigna, MetLife, United Concordia, BCBS, Aetna, Humana, and most PPO and HMO carriers), and surfaces remaining annual maximum, deductible status, frequency limits, and coverage percentage on common codes (D0150, D0210, D1110, D2740, D2950, D4341, D7140). The front desk sees it on one screen the morning of.
Why it matters. The biggest source of front-desk firefighting is the surprise bill. When the hygienist finishes a prophy and the patient sees the receipt at checkout, an unverified coverage gap turns a $0 visit into a $312 visit. The patient blames the practice. The five-star review becomes a two-star review. Treatment plan acceptance on the crown we just diagnosed drops from 62 percent to 19 percent.
Why the front desk cannot keep up manually. Verification takes 7 to 12 minutes per patient on the carrier portal. A 3-chair practice with 22 patients on tomorrow's schedule needs 3 to 4 hours per day on verification alone. They do not have those hours. They skip it. The surprise bill happens. The cycle repeats.
The mechanism
- Capture. AI voice agent (module 02) or front desk enters carrier and member ID during booking. No additional patient-facing work.
- Real-time verify. Eligibility check fires the night before. Pulls 270/271 EDI or carrier API. Most carriers respond within 90 seconds.
- Benefits parse. Annual max, deductible used/remaining, frequency limits on prophy and FMX, age restrictions on sealants and fluoride, downgrade clauses, coinsurance percentage by procedure category.
- Front-desk display. Single line per patient in the morning huddle: "Maria, Delta PPO, $1,250 remaining, prophy 100%, restorative 80% after $50 deductible, no FMX this calendar year."
- Prior auth flagging. System flags procedures that typically require prior auth (crown buildups, SRP per quadrant, oral surgery) 72 hours out so the front desk can submit before the chair.
- Treatment plan tie-in. When the doctor diagnoses chairside, the assistant pulls real-time remaining benefits to present an accurate same-day estimate. The illustrative target is a 28 to 42 percent lift in acceptance.
Sample booking-call insertion
Illustrative target
No-Show Recovery Sequence
What it does. When a patient no-shows or cancels with less than 24 hours notice, the open block triggers a recovery sequence. SMS goes to the standby list first (patients who asked to be "called if anything opens"). If standby does not fill, the system pulls patients from the next 21 days of recall list whose visit can move up. If still unfilled, the block goes to patients with pending treatment plans whose benefits expire at year-end.
Why it matters. Illustrative math: at an assumed $180 to $340 in production per chair hour, an assumed 14 percent no-show rate on a 3-chair, 8-hour-a-day practice is ~33 vacant hours per month, or $7,200 to $11,200 in production walking out the door. Most practices accept this as a tax. Part of it is recoverable: in a 2026 meta-analysis of 10 trials of hospital outpatient appointments, patients who got reminders were on average 11 percent more likely to attend than patients who got none.
Why the standby list alone is not enough. Most practices keep the standby list on a sticky note. When the cancellation comes in at 8:47am for an 11am block, the office manager has 2 hours to call 14 people, reach maybe 4, and book 1. The AI hits all 14 in parallel via SMS in 90 seconds and locks the block from whoever taps yes first.
The mechanism
- Trigger. PMS sync detects the canceled or no-showed block within 5 minutes. Sequence fires.
- Wave 1, standby. SMS to standby patients matching procedure type and time window. First "yes" books. Rest get "thanks, we'll keep you on the list" within 30 seconds.
- Wave 2, pull-up. If standby does not fill in 30 minutes, system identifies patients booked in the next 14 days who could move up. SMS offers the earlier slot.
- Wave 3, pending treatment. If still unfilled, surface patients with diagnosed but unscheduled treatment plans whose benefits expire in 60 days. AI voice call offers to start now.
- Confirmation cascade. Every booked appointment gets a 48-hour reminder, 24-hour confirm-or-cancel SMS, and 2-hour day-of nudge. A second touch helps the riskiest bookings: in a Penn Medicine trial published in NEJM Catalyst in 2026, adding an automated call to text reminders cut no-shows among high-risk patients to 9.6 percent from 11.3 percent.
- Cancellation fee policy. If the practice has one, system enforces consistently. Fee disclosures SMS-confirmed on booking. The disclosed policy is the goal, not the fee.
Sample no-show recovery SMS sequence
Illustrative target
Review Acceleration Engine
What it does. Sends a single, well-timed SMS within 90 minutes of the patient leaving the chair with a one-tap link to your Google review form. Every patient gets the same request and the same link, whatever their experience. Any reply that raises a concern routes to the office manager's inbox with the patient's comment, so the problem gets a personal follow-up.
Why it matters. Google says prominence, one of the factors behind local results, is based partly on how many reviews you have, and that more reviews and positive ratings can help your local ranking. Take two hypothetical practices: one with 312 reviews at 4.8 stars, one with 47 at 4.9. The first has far more of the review volume Google describes. Patients look at freshness too: in BrightLocal's 2026 survey, 74 percent of consumers look for reviews written in the last three months. The illustrative target here is 1.2 reviews per quarter to 14. Treat Maps rank as a result to track, not a promise.
Why review software alone is not enough. Standard review software sends one static template and gets ignored. The Nirvani engine times the request to the sentiment window (60 to 120 minutes post-visit is the peak), filters out patients who recently reviewed, and rotates copy variants to prevent SMS fatigue. It never screens patients by how their visit went.
The mechanism
- Trigger. PMS flags the appointment "completed" at checkout. Sequence fires 75 to 105 minutes later (randomized for natural feel).
- Same ask for everyone. One SMS with the Google review link goes to every eligible patient, whatever their experience. It also invites a reply with any feedback.
- Review handoff. Direct link to your Google Business Profile review form. One tap and the patient is in. In BrightLocal's 2026 survey, 28 percent of people said they will always write a review if asked.
- Concern follow-up. Any reply that raises a concern gets a personal SMS from the office manager within 10 minutes ("I'd like to make this right. Can we talk?"). The patient keeps the same review link, and no one is asked to hold off on a review.
- Suppression. Recent reviewers, do-not-contact list, and patients under 18 are excluded automatically. No one is excluded based on how their visit went.
- Reporting. Daily review velocity dashboard shows new reviews, sentiment, and Maps 3-pack rank tracking for top local terms.
Sample post-visit SMS
Illustrative target
09. Roadmap
The 30/60/90 implementation roadmap.
Three phases. Each has a verifiable dashboard milestone before the next starts. We do not move forward if the prior phase is not green. This is the rhythm built to get a practice fully live inside 90 days without breaking the front desk.
Connect, configure, ship Module 01.
- PMS integration (Dentrix, Eaglesoft, Open Dental, Curve, Denticon) live.
- Practice phone number ported or forwarded. 24/7 ring routing configured.
- A2P 10DLC SMS campaign registered and approved (review can take 1 to 4 weeks, per The Campaign Registry).
- Patient consent backfill workflow shipped to front desk.
- Module 01 (Hygiene Recall Engine) deployed and monitored, with SMS steps live once the campaign is approved.
- BAA executed and stored.
Modules 02, 03, 04 live.
- Module 02 (24/7 New-Patient Intake) goes live with custom intake script.
- Module 03 (Eligibility Pre-Screen) connected to top 5 carriers by patient volume.
- Module 04 (No-Show Recovery) live with standby list and confirm cascade.
- Front-desk huddle dashboard fully populated and trained.
- Target: first measurable lift in recall reactivation rate and new-patient bookings.
- Weekly check-in with deployment lead.
Module 05 live. Stack tuned.
- Module 05 (Review Acceleration) live. Target: first review velocity bump.
- Eligibility coverage expanded to remaining carriers in the patient base.
- Script A/B tests locked in. Best-performing variants live.
- Treatment-plan acceptance dashboard launched.
- Dashboard handoff complete. Practice operates the stack day-to-day without Nirvani in the room.
- Monthly executive review cadence begins.
Timelines are targets, not guarantees. The plan builds in buffer for PMS integration and A2P campaign review.
10. ROI Snapshot
What the math looks like per chair.
Illustrative model: a 3-chair, $1.5M general practice on a scoped Nirvani engagement. A modeling assumption, not a measured result. Numbers below are net of platform cost.
Multi-location or 6+ chair practices run on Nirvani Multi-location engagement, which adds multi-location PMS sync, location-level rollups, and a dedicated deployment lead. Per-chair ROI math is the same, multiplied across more chairs.
11. Illustrative Example
A fictional 3-chair practice. Before and after.
Illustrative example. This is a fictional practice built to show the math, not a real customer. Every figure below is modeled from the assumptions on this page, not measured.
The 90-day arc.
In this modeled example, the practice starts in February with a 36 percent recall rate, 13.5 percent no-show rate, and 84 reviews on Google at 4.7 stars (most 18 months old). Two front-desk members are stretched thin. The owner is 7 weeks from quoting a fourth hire. The stack ships in sequence. Module 01 goes live on day 11 and produces its first re-booked hygiene visit on day 14. Module 02 catches 23 after-hours new-patient calls in its first 10 nights, 19 of which book. Module 03 heads off 4 surprise-bill conversations in the first 30 days. Module 04 recovers 31 of 44 cancellations in its first month. Module 05 ships day 67. By day 90, the model shows 41 new Google reviews at 4.9 and the fourth-hire search called off.
- Recall reactivation rate: 36%
- New-patient bookings: 28 / month
- After-hours call recovery: 4%
- Eligibility verified on arrival: 29%
- No-show rate: 13.5%
- Google reviews this quarter: 1
- Avg star rating: 4.7
- Owner admin hours / week: 18
- Recall reactivation rate: 79%
- New-patient bookings: 47 / month
- After-hours call recovery: 91%
- Eligibility verified on arrival: 97%
- No-show rate: 3.1% effective
- Google reviews this quarter: 41
- Avg star rating: 4.9
- Owner admin hours / week: 4
In the model, net production lift at day 90 is $34,200 per month. In this fictional example, the owner reinvests in a new digital scanner ($28K), cuts clinical days from 4 to 3, and uses the freed time to start onboarding a second associate. The fourth front-desk hire is never made.
12. Run the Numbers
Validate the math with your inputs.
Before you book a call, run two calculators with your own numbers. Both default to dental practice ranges and produce a defensible projection in under 60 seconds. No signup required.
Voice Agent ROI
How much revenue your voicemail is losing you. Set your industry to "Dental Practice." See the after-hours leakage in dollars.
LTV Calculator
Lifetime value per new patient. Use this to set defensible CAC ceilings for your local SEO, Google Ads, and referral programs.
Stack Assessment
Book a 30-minute call with the founder.
We walk through your current numbers (recall rate, no-show rate, after-hours call leakage), map them to the stack modules, and give you a written deployment plan with realistic timeline and modeled lift. No sales pitch. No second call with a closer. If the math does not work for your practice, we will tell you on the call. If it does, we ship.
Meeting14. FAQ
The questions we get on every assessment call.
Compliance, integration, training, contracts, and the question every owner is too polite to ask first.
Is the stack HIPAA-compliant?
HIPAA compliance is a posture, not a checkbox. It depends on how your team uses the system. Run our BAA documentation and security attestation package by your compliance counsel before deployment. We are not your legal advisors and will not pretend to be.
Which practice management systems do you integrate with?
What does training look like for the front desk?
What if our front desk thinks we are trying to replace them?
The aim is a front desk that can take a real lunch, with a phone that is no longer on fire all day.