OrbDoc

Ambient capture and structured documentation are the whole product. The speech layer is licensed rather than built, running on Deepgram's Medical Nova 3, with Google Cloud AI and the vendor's own medical algorithms above it, which is the same shape as Sayvant building on Azure OpenAI and does not displace the model as the artifact. Naming the dependency rather than obscuring it is what separates this from the licensed whole product cases graded C.

The review gate is stated plainly and consistently across materials: all notes are presented for provider review before EHR submission, and the vendor describes the only change to workflow as a quick review step before finalising. Approval precedes the write rather than following it, which is the safer of the two orderings this index tracks. The audit trail supports it by recording when notes were created or modified and by whom. Held at B because no acceptance rate, edit burden figure or confidence threshold is published.

The most specific supply chain disclosure located in this wave. OrbDoc names its speech model, Deepgram Medical Nova 3, its cloud AI layer, Google Cloud, and acknowledges its own proprietary algorithms on top, which is the third answer in 65 records to the question of whose model is running and by far the most precise.

It also names HL7 FHIR as its interoperability standard rather than claiming integration generically, and its audit trail records which AI models were used for each documentation action, so model provenance is reconstructable after the fact. Held at B because no accuracy figure, model card, error rate or evaluation methodology was located.

The most precise supply chain disclosure located in this lane, and it does something no other record here does. The vendor names its speech model by product and version, names the cloud artificial intelligence layer it runs on, and acknowledges its own proprietary algorithms sitting on top, so a buyer can see all three layers of the stack rather than being told the technology is proprietary.

Naming a speech model specifically matters more in this category than naming a language model, because speech recognition is where accent, dialect and impairment cause the errors that fall unevenly across patients, and a named model has published characteristics a buyer can research. The interoperability standard is named rather than described as integration generically. What lifts this above the other top graded records is the audit trail.

It records which models were used for each documentation action, so model provenance is reconstructable for a specific note after the fact. A static list tells a buyer who is in the chain today; a per action record tells them which system produced the note in front of them, which is what actually matters when a particular document is disputed months later.

Ask two follow ups rather than treating the disclosure as complete: whether a sub processor list exists covering parties beyond the model layer, and whether customers are notified when a model in that stack is changed.

Better structured than most vendor claims and still anonymous. A cardiology group case study reports 45 minutes of daily time saved per provider, a 23 percent improvement in RVU productivity, and physician satisfaction rising from a 2.8 baseline to 4.2 out of 5 after six months, and quoting a BASELINE and a follow up interval is methodologically ahead of the bare percentages typical here. A telehealth deployment across 35 states with more than 800 contracted providers is also described. But no customer is named, no third party verified any of it, and no study or controlled evaluation exists, so this is well presented vendor reporting rather than evidence.

This record needs reading carefully, because the vendor's own sources disagree with each other.

A strong data position appears on OrbDoc's per EHR integration landing pages: audio processed in real time and immediately deleted with no persistent storage, Google Cloud HIPAA infrastructure, full business associate agreement coverage, and text notes retained under the customer organisation's own retention policy. The vendor's own FAQ, which is the page a clinician actually reads, says something much thinner. Data is protected with encryption in transit and at rest, access is limited to authenticated workflows, and personal information is not sold. It states no retention schedule, no audio deletion commitment and nothing about training use. A third party listicle separately attributes seven years of audio retention to this product, which is the opposite of the immediate deletion the landing pages claim.

Encryption in transit and at rest plus a no sale statement is a real but ordinary position, and that is a C. Ask for the audio retention answer in writing, because three sources currently give three different ones.

The integration landing pages state that the product is fully HIPAA compliant with enterprise grade security and that full business associate agreement coverage is provided, and the integrations page says BAA documentation is provided for covered implementations.

But the vendor's own FAQ asks the question directly, Is OrbDoc HIPAA compliant, and does not answer it. The reply is that OrbDoc is built for healthcare documentation workflows and that the reader should review the privacy policy and terms of use for the current public terms. Declining to say yes in the one place a clinician goes to ask is materially different from saying yes in marketing copy, and the more careful language is almost always the more accurate one.

This is not a finding that the product is non compliant. It is a finding that the vendor's positions differ by page, and for a product recording patient encounters that is the first thing to settle in writing before use.

Two claims that need separating. Data is stored in Google Cloud HIPAA compliant infrastructure, which certifies the platform OrbDoc runs on rather than OrbDoc's own controls, the same attributed certification pattern this index tracks for RXNT and CareCortex. Regular security audits and compliance reviews are described as process rather than as a named, dated report.

JCI and CBAHI are healthcare facility accreditation standards rather than software attestations, so compliance with them is doing loose work in a software context. No SOC 2, ISO 27001 or equivalent naming OrbDoc itself was located.

No clearance claimed and none required for ambient documentation, and no United States device pathway attaches to a draft note the clinician reviews and finalises. The vendor is explicit that a review step before finalising is the only addition to the existing workflow, which puts the signature exactly where the regulatory reasoning needs it.

This record is one of the narrower ones in the category, and that is worth stating rather than recorded as a gap. The scope expansion pattern documented across most ambient scribes in this index, meaning coding engines, real time clinical insight, chart question answering, patient facing products and revenue cycle modules, is largely absent here. The closest approach is documentation described as linking evidence to support medical decision making and procedure coding, which is a claim about producing a defensible record rather than about proposing codes. That distinction carries weight: supporting the coding of what happened is a different object from optimising toward the highest justified level. Nothing located shows a decision support, patient facing or reimbursement module.

Following the treatment this index gave Nabla, the narrowness is credited rather than read as immaturity. It keeps the regulatory position simple and it is a commercial choice the rest of the category has mostly declined to make.

Markets appear to be United States only. The named integrations are United States record systems, the customer language is emergency departments, integrated delivery networks and academic medical centres, and no United Kingdom or European Union presence was located. If that changes, ambient documentation is treated as software as a medical device in those markets and this position would need reassessing.

Graded up on a mechanism rather than a policy document. The audit trail records which AI models were used for each documentation action alongside who accessed what and when, which means a health system can reconstruct after the fact what produced a given note and on which model version. That is model provenance at the record level, and it is the single most useful governance artifact found in this wave, because every other vendor's audit story stops at user actions.

Separately the vendor states its diarisation is trained on diverse accents and speech patterns common in healthcare settings, an affirmative claim on the failure mode this index tracks across speech products.

Held at B because the accent claim is asserted rather than demonstrated, no subgroup or fairness data is published, and the RVU productivity framing in its case study is revenue adjacent even though it is presented as productivity rather than coding intensity.

One mechanism here is genuinely strong for after the fact investigation, and it is the reason for the grade. The audit trail records which models were used for each documentation action, so when a note is disputed months later an organisation can establish what produced it rather than reasoning about the product in general. That is the evidence any correction, complaint or internal review would need, and almost nothing else in this lane provides it.

Against that, the vendor's published statements contradict each other on a basic question, which is a problem in its own right rather than merely a gap. Per integration landing pages state that audio is processed in real time and immediately deleted with no persistent storage. The frequently asked questions page, which is what a clinician actually reads, states no retention schedule and no audio deletion commitment at all.

A third party listing separately attributes years of audio retention to the product. Three sources, three answers, and a buyer cannot rely on any of them. Where a vendor's own pages disagree, the operative version is whichever one ends up in the contract, so that is where the question belongs. No accuracy or error figure, no published limitations and no warranty, indemnity or remediation commitment was located. Ask for the retention answer in writing and for an accuracy figure with a definition.

Graded on a direct contradiction inside the vendor's own site.

The integration landing pages describe named, standards based, genuinely bidirectional exchange over HL7 FHIR across clinical, administrative and revenue cycle modules for Allscripts Sunrise Acute Care, TouchWorks, Professional and Veradigm, supporting both cloud hosted and on premise deployments, plus direct audio capture from Zoom for Healthcare, Microsoft Teams for Healthcare and Doxy.me. The vendor's own FAQ says OrbDoc is currently available on iOS devices, iPhone and iPad. The App Store listing describes export to major EHR systems including Epic, Cerner and Athena.

An iOS only application does not perform bidirectional exchange with an on premise Allscripts Sunrise Acute Care instance, and export is not write back. Enzo Health was marked down on the same pattern, claiming both integration and replacement while naming no system for either. The capability may well exist for enterprise implementations, but it cannot be graded from marketing pages that the vendor's own support documentation contradicts. Establish in a live sandbox which mechanism actually runs for your EHR before treating this as integrated.

Graded for the same sourcing reason as the EHR and PHI axes on this record.

The specifics that appear on the integration and regional landing pages, Google Cloud HIPAA infrastructure named as the hosting platform, audio not persisted, support for mixed connectivity and on premise record systems, and a Gulf regional offering implying regional deployment, do not appear in product documentation. The vendor's own FAQ states the product is currently available on iOS devices only and says nothing about hosting region, residency, sub processors or offline operation.

A separate third party listicle credits OrbDoc with an offline capable architecture that records and generates notes with no connectivity, which would be a genuine differentiator in this lane alongside voize, Empathia, Enzo and Aegira, but that capability appears nowhere in the vendor's own materials and should not be relied on without a written confirmation or a test with the device in airplane mode.

A free trial is offered and promoted prominently, and a rural hospital testimonial describes the product as advanced documentation at an accessible cost, but no rate card, tier structure or per provider figure was located for any market. For a vendor selling simultaneously to integrated delivery networks and to rural single site hospitals, the absence of any published anchor is a larger gap than usual.

Wide across settings and, uniquely in this index, across regions. Coverage spans emergency departments, inpatient services, ambulatory clinics, cardiology, paediatrics, family medicine and mental health, sold to integrated delivery networks and academic medical centres at one end and to rural hospitals, small practices and community care teams at the other, with cross device continuity so a note started on a phone during rounds is finished on a tablet at the nurses' station.

The distinguishing element is geographic: a dedicated Gulf offering with native Arabic support and alignment to JCI and CBAHI standards across Saudi Arabia, the United Arab Emirates, Qatar, Kuwait, Bahrain and Oman. No other vendor assessed in this wave serves the Middle East at all, and Arabic clinical documentation is a genuinely different problem from adding a language to a menu.