Does DeepScribe Work Inside athenahealth?

On athenaOne's published interfaces, by DeepScribe's account: the FHIR R4 API with OAuth 2.0 for the clinical read and the proprietary athenaOne APIs for encounter document sections and the schedule. Reading the chart before the visit is the part that distinguishes DeepScribe's description from most competitors on this platform, and it is the part to verify first. A scribe that knows the problem list and medication list before the conversation starts produces a different note from one that infers them from audio.

Discrete field write back is the second claim. DeepScribe says it routes content into specific athenaOne sections rather than dropping a narrative into one field, and that coding suggestions land in the workflow. Whether diagnoses are written as structured entries on athenaOne diagnosis cards, and whether E/M level suggestions are visible to the biller inside athenaOne, is not published at field level. Not published. Checked September 2026.

The incumbent is Ambient Notes. DeepScribe's counter is specialty depth: it names orthopaedics, cardiology, neurology, gastroenterology, urology and oncology among others, and reports a KLAS score of 98.8 and a 95.9 percent major defect free rate. Both figures are DeepScribe's citations of KLAS research and its own quality programme respectively.

Before the visit, it loads the patient's context from athenaOne. During the visit, it records with consent. After the visit, it produces a specialty formatted note and pushes it into the athenaOne encounter sections, with coding suggestions, for the clinician to review and sign in athenaOne. DeepScribe positions the specialty formatting as the product: a long neurology history of present illness or an orthopaedic procedure note is structured the way that specialty expects, rather than as a generic SOAP note.

For groups carrying risk, the HCC capture is the financial argument. Treat it carefully. A suggested HCC that a clinician accepts without review is a compliance exposure, not a revenue win, and the coding automation page covers the oversight policy that has to sit beside the feature. The charge capture leakage calculator will tell you whether your current under coding is large enough to justify a coding oriented scribe over a cheaper one.

The ambient scribe use case page lists the write back questions we put to every vendor, and they apply here with one addition: ask to see the athenaOne encounter after a test push, not a screenshot of DeepScribe.

Everything, because DeepScribe is a direct purchase. Your practice authorises the application against your athenaOne data, signs a BAA directly with DeepScribe, agrees which clinicians and specialties go first, and owns the relationship when something fails. If DeepScribe's Marketplace listing is current, athenahealth's technical and security review applies at the time of that review. It does not cover retention, subprocessors, model training on your encounters or breach notification, which are your contract terms. The HIPAA and AI page sets out the minimum.

Because the product reads the chart before the visit, ask what is retained from that read and for how long. A scribe that stores a copy of the problem list and medication list for every scheduled patient, including no shows, holds more PHI than one that only stores the encounter. The scribe compliance checklist puts this and the other questions in order.

DeepScribe is sales led with no self serve signup or free trial. Plan for a procurement, not a download, and set a decision date before the first demo.

• Marketplace listing unconfirmed. We could not verify a current DeepScribe listing on the athenahealth Marketplace in September 2026. Ask for the link.
• No published pricing, no trial. DeepScribe prices on request. Not published. Checked September 2026. The DeepScribe pricing page covers what to ask for.
• Field level write back not documented publicly. Discrete field routing is claimed; the specific fields and behaviour when they already contain text are not published.
• Outside Ambient Notes. Separate contract, separate security review, competing from outside the encounter against embedded models.
• Coding features carry compliance obligations. HCC and E/M suggestions need a review policy and an audit trail on your side.
• Pre visit chart read expands the data footprint. Confirm what is retained for patients who are scheduled but not seen.

Ask, in writing: the athenahealth Marketplace listing link and last review date; which athenaOne interfaces and scopes the app uses; exactly which chart data is read before the visit and how long it is retained; which fields are written and what happens when they already contain text; whether diagnoses and E/M levels land as structured entries visible to billing; how a failed write is surfaced; where audio and transcripts are stored and for how long; whether your data trains shared models; the price per clinician and minimum term; and the offboarding process. The EHR integration question list is the full set.

Ask for references on athenaOne in your specialty specifically. DeepScribe's strongest references are in oncology and specialty systems; an athenaOne primary care reference tells you more than an OncoEMR one.

Then trial it against one Ambient Notes model on the same real encounters. The scribe comparison sets out what to score, and for specialty groups the coding accuracy column matters as much as the note.

DeepScribe's athenaOne claims are among the strongest on this platform and among the least independently documented. Verifying them takes a test patient, an inspection of the encounter afterwards and someone who has seen how the embedded alternatives behave. Clunic takes no vendor commissions. Our vendor selection engagement runs that verification, compares DeepScribe against the Ambient Notes models on your own encounters, and scores the contract terms. If coding oversight is the concern, AI governance and compliance builds the review policy first. Book a call to start.