There is a moment in every technology cycle when the leaders of one wave realize they need to move into the next one. They have customers, capital, and credibility. What they have to figure out is architecture.
That moment is happening right now in dental AI. And the decisions being made, quietly, inside Pearl and Overjet tell you a lot about where this industry is going. They also tell you exactly why we built OraCore the way we did.
The Paths Taken
Pearl and Overjet built their reputations on the same core capability: reading dental images. Feed them an X-ray and they return findings, flags, and probabilities. It is genuinely impressive technology. It is also a fundamentally different problem than what a scribe does.
Reading a dental image is a computer vision task. The input is a file. The output is a structured set of labels. The model is trained on annotated radiographs. The environment is controlled.
Ambient audio capture is none of those things. The input is a live acoustic environment full of suction motors, handpieces, patient breathing, hallway conversations, and six different voices talking across each other. The output needs to be a compliant, permanent clinical record. The environment is chaotic. Getting it right requires a different architecture, different training data, different audio engineering, and a different product philosophy.
Pearl recognized this gap and built their scribe capability directly on top of their existing imaging platform. Overjet took a different route: they acquired DentalBee in late 2025 and integrated a scribe that was already built. Two companies. Two paths. Both arriving at the same realization that the documentation problem is where the next battle is being fought.
Neither path is wrong. But the route you take shapes the product you end up with.
The Architecture Problem
The scribe problem is not an interface problem. It is an architecture problem. And the place where that becomes most clear is in how audio is captured and processed.
A generic voice-to-text workflow is straightforward. You start recording, the model converts speech to text, you stop recording. The cost is low because the approach is simple. But simple has consequences in a dental operatory.
When a scribe captures audio in fragments, or converts to text in real time without full context, it misses things. It mishears words. It loses clinical nuance. It produces a transcript that reads like a transcript, not like a clinical note written by someone who understood what was happening in that room. In a casual conversation, that is fine. In a permanent legal and clinical record, it is not.
OraCore captures the full appointment audio and transcribes after the visit, not in fragments during it. That means our AI has the complete clinical context before it generates a single word of the note. The output is cleaner, more compliant, and more complete. We made that architectural decision knowing it costs more to build and operate. We made it anyway, because we do not think a clinical record should be built on a technology shortcut.
Here is the part that surprises most people: despite spending more on the underlying technology, OraCore is priced lower than most of the alternatives.
That is not a pricing anomaly. It is a deliberate choice. Better technology should not mean a larger bill. It should mean more practices can actually afford to use it. Per-seat, enterprise-contract pricing is not how independent dental practices operate. So we did not build that way.
Operatory Audio Is a Specialized Problem
A dental operatory is not a conference room. It is not a doctor’s office. It is a specific acoustic environment with specific noise characteristics, specific workflows, and specific terminology that a general-purpose AI system simply does not understand.
Every single OraCore account is customized to the practice. Not in the sense of checking a few boxes in a settings screen. Customized in the sense that we build around how your team actually works.
Use a DIAGNOdent or an iTero scanner? OraCore hears the workflow, recognizes the context, and gets the documentation right. Have a provider name like Hutchison, not Hutchinson? We have you covered. Use a shorthand for a particular treatment approach that is specific to your practice? We learn it.
Beyond catching specific words and names, there is a HIPAA dimension that most scribe companies do not address seriously. Because OraCore listens to the full visit before transcribing, our AI is trained to identify and remove conversations that filter in from the hallway or an adjacent operatory. That is not a feature bolted on after the fact. It is a consequence of taking the full-visit capture approach seriously from the start.
We also offer support for a second “negative microphone,” a second mic placed specifically to capture and cancel a noise source rather than the clinical conversation. The system subtracts the noise-only signal, which isolates voice with dramatically higher precision. For orthodontic offices with open-plan layouts, or any practice that plays background music, this is the difference between a scribe that works and one that frustrates.
The operatory is not a controlled environment. Our architecture treats it that way.
Seats Are a Tax on Growth
Dental practice staffing does not look like a software company’s org chart. It changes. Hygienists take maternity leave. Dentists go on vacation. Specialists come in for one day. Temps cover gaps. An associate starts, tries the practice, and moves on.
Seat-based pricing ignores all of this. A scribe tool priced per provider can easily reach 800 dollars or more per month before you have accounted for any of those scenarios. You pay when someone is out for three weeks. You pay when a traveling specialist comes in for one oral surgery day. You pay when you hire someone who does not work out.
That is not how dentistry works, and it should not be how dental software works.
I have been in this industry for a long time. The thing that always frustrated me about software pricing was when I had to remember to manage it. When I found myself holding back from adding a useful workstation because IT was going to charge me another 150 dollars a month to service it. When a staff change meant a billing change I had to remember to make. When the software was billing me for someone who had not logged in for two months.
I will not build that product.
With OraCore, if someone is out on vacation, you are not paying for them. Bring in a new employee? Add them. Let your team have their own accounts, with their own preferences, their own workflow customizations. You are not managing headcount in a pricing dashboard. You are running a practice.
Our pricing is based on hours of operatory time captured, not seats. The hours pool across your entire team. Add a provider for a single day and they are covered under the same plan. Slow month? You use fewer hours. That is it.
That is what a team workflow looks like. And that is the pricing model dentistry actually deserves.
What This Means for Your Practice
Pearl and Overjet are building toward where the market is going. They have resources, and they have imaging moats that are genuinely valuable. But imaging AI and ambient documentation AI are different problems, requiring different architectures and different teams to get right.
OraCore was built for the documentation problem from day one. Full-visit capture. Dental-native AI. True customization at the practice level. Pricing that does not punish you for how dentistry actually works.
The question is not whether AI is coming to the dental operatory. It already arrived. The question is whether the tool you adopt was built for your practice, or built for someone else’s and adapted to fit.
We built OraCore for yours.
Ready to see how it works in an operatory like yours?
Book a demo and we will walk you through what full-visit capture looks like for your specific setup, your PMS, and your team.
Frequently Asked Questions
Image-reading AI focuses on analyzing static visual data like X-rays in a controlled digital format. Ambient audio capture must process a live, chaotic acoustic environment with background noise, overlapping voices, and clinical context to generate a compliant dental record. The architectural requirements are fundamentally different.
Capturing the full visit rather than recording in short fragments gives our AI the complete clinical context before generating the note. This prevents missed nuance, improves accuracy, and allows our system to filter out background hallway conversations to better protect HIPAA compliance.
Traditional scribes charge a flat fee per user, penalizing practices for part-time staff, vacations, or adding team members. OraCore charges based on hours of operatory time captured, pooling across your entire team so you only pay for what you use.