Archy

A dental practice management system adding an AI layer, and the platform is the purchase. The core product covers charting, imaging, scheduling, insurance, payments, payroll, and patient communication, sold on the proposition that one system replaces a stack of add on tools. The AI is real and increasingly native, notably a scribe and voice perio charting, but a practice buys Archy to replace its practice management software and inherits the AI. The imaging AI is a third party dental detection product embedded through partnership rather than built in house.

Clinician review is the stated model across the AI surface. The scribe produces a complete clinical note as a SOAP note or mapped to practice templates for the clinician to accept, voice charting records spoken perio findings into the record, and the embedded imaging AI is positioned explicitly as a second opinion during treatment conversations rather than as a diagnostic authority. Framing detection AI as a second opinion is the appropriate posture for software operating alongside a dentist reading the same radiograph.

One genuinely useful architectural disclosure and little else. The company is specific that its scribe runs natively inside the practice management system with direct access to the patient chart, treatment history, and planned procedures before recording begins, in contrast to standalone scribes connected through integrations, which is a concrete and verifiable design claim rather than marketing. No model provider, training data, or accuracy measurement is published for any of its own AI features.

Nothing published identifies any party in the chain. No foundation model provider, model class or version is named, no training data source is described, no hosting arrangement is published for the artificial intelligence features, and no sub processor list was located in two passes. The architecture disclosure that does exist works against the vendor on this axis rather than for it.

The scribe is stated to run natively inside the practice management system with direct access to the patient chart, treatment history and planned procedures before recording begins, which is a concrete and verifiable design claim and a genuine differentiator against scribes connected through integrations.

It also means the content available to whatever systems sit behind that feature is the whole record rather than one conversation, so the unnamed chain has a wider reach here than for a standalone scribe. One further data location detail belongs in the same conversation. An offline imaging mode stores radiographs locally and uploads them when a connection returns, so clinical images exist on practice hardware the vendor does not manage as well as in the platform. Establish what protects the local copy, whether it is removed after upload, and who is responsible for it. Ask for a sub processor list, and ask what the ambient feature can read from the chart.

No quantified evidence was located for the AI features. Claims are qualitative and forward looking, describing reduced after hours charting and faster documentation, supported by named dentist testimonials rather than measurement.

The company also describes portions of Archy Intelligence as rolling out progressively with further agents in development, so some capability is roadmap rather than shipped. That is a fair thing for a company to do and it says so openly, but it means a buyer evaluating the AI layer should establish which components are live in the version they would receive today rather than which are described on the site.

No retention period, no statement on whether customer data is used to train or improve models, and no de identification posture was located.

The questions are live because of what the platform holds and what the AI layer added to it. As a cloud practice management system it holds the complete patient record including radiographs, and the scribe adds audio capture of the clinical encounter. Retention of those recordings, and whether they or the chart data improve future models, are both unanswered.

One published architectural detail belongs in the diligence conversation. An offline imaging mode stores radiographs locally and uploads them once a connection returns, so images exist on practice hardware as well as in the platform. A practice should establish what happens to the local copy, how it is protected on a machine the vendor does not manage, and whether it is removed after upload.

No business associate agreement terms and no HIPAA statement addressing the company's own obligations were located. Business associate status is structurally certain for a cloud practice management vendor holding patient records, and agreements exist in customer contracts, but nothing is published.

The company describes HIPAA compliant infrastructure and HIPAA compliant chat. Those describe an environment and a feature rather than the contractual instrument. A practice needs to know what the agreement says about breach notification timing, subcontractors, permitted uses and return or destruction of data on termination, and none of that is knowable before entering a sales process.

No SOC 2, ISO 27001 or HITRUST attestation of the company's own was located, and no trust centre was found.

What the company does publish is that the platform runs on HIPAA compliant Amazon Web Services infrastructure across multiple data centres with automatic backups. That is the cloud provider's compliance posture rather than the vendor's own. Amazon will sign a business associate agreement and offers HIPAA eligible services, which makes a compliant application possible. It does not make the application compliant, and it says nothing about how this vendor configures, monitors or restricts access to what it stores there.

A third party listing site attributes ISO 9001 and SSAE 18 to the product. Neither is credited, on two separate grounds. It is not the vendor's own publication, and aggregator listings are not treated as evidence of certification in this index. And both are mischaracterised in any case: ISO 9001 is a quality management standard rather than an information security one, and SSAE 18 is the attestation standard under which a SOC report is issued rather than a certification any organisation holds.

For a cloud practice management system holding the complete dental record including radiographs, an attestation of its own and its stated scope should be the first item a practice requests.

The practice management platform and the scribe carry no FDA pathway, correctly, since documentation and administration are outside device regulation. The distinction worth drawing for buyers is that the diagnostic imaging AI available inside Archy is FDA cleared dental detection software supplied by a third party partner, so the clearance belongs to that partner's product rather than to Archy. A practice relying on radiographic detection is relying on the partner's regulatory standing, not the platform's.

No AI governance framework or bias evaluation was located, for either the native AI features or the imaging AI embedded through a third party partnership.

The partnership is the part worth separating out. Clearance held by the imaging AI vendor covers that vendor's device, and it does not describe how the model behaves once embedded in a different company's practice management system, with that system's image handling and that system's user interface deciding how a finding is surfaced to a dentist. The buyer of the platform is not the holder of the clearance, and no published material addresses monitoring of the combined product as deployed.

Two passes located no accuracy or error figure for any of the artificial intelligence features, no published limitations, no warranty, indemnity or remediation commitment. The one substantive published claim is architectural rather than performance related, that the scribe reads the chart, treatment history and planned procedures before recording begins, and it raises the stakes of an error rather than lowering them.

A system with prior chart context can produce documentation that is confidently wrong in a way a transcription only product cannot, because it can carry a stale or mistaken element forward from the record into a new note where it acquires the appearance of fresh clinical observation.

Nothing published describes how that is guarded against, what the clinician sees about the provenance of a given statement in the draft, or whether chart derived content is distinguished from what was actually said in the room. The buyer here is typically a small practice without a compliance function, which changes what a lack of published commitments means in practice.

A large health system can negotiate representations into a contract; a single site dental practice signs the standard terms, and the standard terms were not located. Ask for accuracy, ask what happens when a note is wrong, and ask specifically whether the draft distinguishes what the system heard from what it read.

The vendor is the practice management system, so integration depth internally is complete: imaging is built in rather than bridged, with native sensor support and TWAIN compatibility plus offline image capture when the connection drops, and the AI features read and write the same patient record. That native posture is the company's central argument against a stitched together stack. The counterweight is that this is a closed vertical system for a single specialty, and no external interoperability standards, HL7, FHIR, or dental data exchange capability were located.

A single cloud platform with a native mobile application for iPhone and Android providing access to schedule, patient messaging, clinical notes, and radiographs. Offline image capture is supported when internet connectivity fails, which is a practical resilience feature for a chairside system. No tenancy, hosting, or data residency terms are published.

Pricing structure is published even though rates are not, which is more than most practice management vendors offer. The company states one platform price per location with unlimited users and unlimited providers included, explicitly rejecting seat based and provider based pricing, with optional Clinical, Front Office, and Marketing suites layered on and special first year pricing for qualifying new practices. A practice can therefore understand how it will be charged and that adding staff will not increase cost, which is the variable that usually surprises buyers, without learning the actual number.

Single specialty by design: general dental practices, from new practice startups through established offices, with functionality spanning the full operational surface of a dental office including payroll, insurance pre authorization submission, imaging, and patient engagement. This is not a multi specialty or medical product, and no hospital, group medical, or dental service organization enterprise positioning was located.