Sources: the American Dental Association's own oral health topic on home care concludes there is merit in tailoring a patient's recall interval to individual need based on assessed risk, and Cochrane Oral Health's review of dental check up frequency found little to no difference between a 6 month and a risk based interval over 4 years, at high certainty. A flat 6 month recall stated as a standard used to sit here and it is not one. Read your own numbers off your own practice management system before you act on anybody's average, including ours.
The three places production leaks out of a dental practice
None of these are clinical problems and none of them are marketing problems. Every one of them is a follow-up problem, which is the only category of business problem that software genuinely fixes.
1. The new patient call nobody answered
A person with a broken molar, or a person who has just started a new job and a new plan, is not shopping carefully. They are working down a list of practices and the list stops at the first one that answers. That call almost never arrives at a convenient moment: it arrives while your one front desk person is on hold with a payer, seating a patient, or gone for the day. A callback the next morning reaches somebody who has already been seen.
The number that matters here is not how many calls you miss. It is what a new patient is worth to you across the first 2 years, counting the exam, the hygiene visits and the treatment that follows. Multiply that by the calls that went to voicemail last month and you have the real figure.
2. The hygiene patient nobody invited back
Hygiene recall is the single highest return automation in dentistry and it is the one most practices never build properly. The usual version is a monthly blast to the whole list, which is not recall, it is a newsletter. Real recall fires on each patient's own due date, names their own last visit, and offers a specific time. A patient who drifts 4 months past due is not disloyal. Nobody asked them.
3. The treatment plan nobody followed up
Diagnosed and unscheduled treatment is production you have already done the clinical work for and never billed. It sits in the practice management system as a list nobody owns. The follow-up that closes it is not a reminder, it is a conversation carrying the cost, what the plan actually covers, and the financing option, because that is where these decisions really stall. The same patient who declined in March will often say yes in September, once their benefit year has reset, and almost nobody asks them twice.
The honest arithmetic: an after-hours answering service runs roughly $140 to $500 a month. Take your own average new patient value over the first 2 years and the service pays for itself on a single family you would otherwise never have heard from. That is why the practices that have done it rarely go back.
What the automation must never be allowed to do
This is medicine, so the limits are not optional and they are not a matter of taste.
- No clinical advice, ever. An intake agent can say what a new patient exam covers and what it costs. It must not tell somebody whether their pain is an abscess, whether to take anything for it, or whether they need an extraction.
- No patient health information on an unsecured channel. If a vendor is going to hold anything protected, that needs a business associate agreement in place first, and the safe build simply keeps clinical detail out of the automated conversation altogether.
- Emergencies escalate to a person. Triage into a same day opening is a scheduling decision, not a clinical one, and anything that sounds like swelling, trauma or bleeding goes to a human immediately.
- It is configured against your protocols, not switched on out of the box. Which plans you take, which providers see which cases, what an emergency slot looks like on your schedule. A vendor default answers none of that correctly.
What this looks like on a real enquiry
Before
A family finds you at eight on a Sunday evening, one of them has a broken molar, and they leave a voicemail. Nobody hears it until Monday afternoon, because Monday morning was 4 hygiene columns and 2 emergencies. By then they have been seen somewhere with a Sunday answering service, and the crown, the exam and the 3 other family members went with them.
After
The same call reaches an AI front desk agent. It confirms you take their plan, explains what an emergency visit covers and what it costs, books the Monday opening inside the same conversation and texts the practice a summary before anybody arrives. They are seated at 9.
The right build for each part of the practice
Hygiene recall and reactivation
- The next visit is scheduled before the patient leaves the chair, because an unscheduled hygiene patient is a lapsed one forming
- Recall fires on each patient's own due date rather than as a monthly blast to the whole list
- Anybody past due gets one specific invitation naming their own last visit, not a newsletter
Unscheduled treatment
- Every diagnosed and unscheduled case is worked as a list rather than remembered by whoever was chairside
- The follow-up carries the cost, the plan coverage and the financing option
- A patient who declined in March is asked again in September, when their benefit year has reset
New patient calls and emergencies
- Every call is answered around the clock, including the plan question that comes before the date question
- Emergencies are triaged into the same day opening rather than left on voicemail
- The practice gets a written summary of who called and why
Orthodontic, implant and full arch consults
- Intake covers the basics and books the consultation in the same conversation
- High-value enquiries reach a human the same hour rather than joining a queue
- The follow-up sequence runs for weeks rather than days, because that is how long these decisions actually take
The tools doing the work
| What it does | Tools | Monthly cost | Setup |
|---|---|---|---|
| AI phone and web intake, answering around the clock and escalating | Smith.ai, Ruby Receptionists, Goodcall, Retell for a custom build | $140 to $500 | Low |
| Patient communication with recall and reactivation built in | Weave, Solutionreach, Lighthouse 360, NexHealth | $200 to $600 | Low |
| Online booking that writes straight into the practice management system | NexHealth, LocalMed, Flex Dental, or the module already inside Open Dental or Dentrix | $100 to $400 | Low |
| Insurance verification before the patient is seated | Zuub, Vyne Trellis | $150 to $500 | Low |
| Review generation after a visit | Birdeye, Podium, NiceJob | $75 to $300 | Low |
| Custom front desk and recall agent across voice, text and your practice management system | Built by OpsJuice on Retell, n8n and Open Dental or Dentrix | Project based | Managed |
Build or buy
Buy first, and buy the cheap thing. Nearly every practice that believes it needs a custom build has a reminder module sitting switched off inside software it already pays for, and a recall report nobody runs. Those cost nothing to turn on and they recover real money inside a month.
A custom build earns its place for a specific reason, not as an upgrade: several plan types needing different answers on the phone, multiple providers with different scheduling rules, or an intake flow your practice management system cannot represent without somebody retyping it into a second screen. If none of those describe you, the off the shelf stack is the correct answer and it is the cheaper one.
The first 30 days, in order
- Day 1. Put an after-hours answering service on your main line and on the number in your listing. Nothing custom, nothing integrated. It recovers the emergencies and the new patients you are losing this weekend.
- Day 7. Turn on appointment reminders and confirmations, and set the failed appointment message to ask for a new date rather than to apologise. This is the cheapest no-show reduction available to a practice and it usually needs no new software at all.
- Day 21. Run hygiene recall off each patient's own due date and work the unscheduled treatment list every week. Recall is the highest return automation in this business and unscheduled treatment is production you have already done the diagnosis for.
Only after those three are running does a custom build make sense. Doing it in the other order is how practices end up paying for an integration that automates a process nobody had agreed on yet.