Avo

The foundation is curated clinical content, not a model. Avo Pathway, described by the company as its foundational product, is a library of digitised society guidelines and calculators, and the platform's stated business model invites content publishers and evidence based organisations to embed their intellectual property into Avo powered copilots. That is the moat is the dataset pattern this index applies to Reveleer and Haystack: strip out the generative layer and a valuable guideline library and no code tool builder remain. The AI is real and does substantial work on top of it, which is why this is B rather than C.

The widest clinical suggestion surface of any product in the scribe portion of this index, with no published limits. Beyond note generation it drafts orders, identifies care gaps, produces differential diagnoses, generates treatment recommendations and suggests DOSING. Dosing in particular is a recommendation with direct patient consequence. No confidence threshold, escalation path, abstention behaviour or acceptance rate was located for any of it. Credit where due on design philosophy: Avo positions itself explicitly against alert driven decision support, arguing for standardising care with love rather than alerts, which is a considered stance on alert fatigue rather than a marketing line. But a philosophy is not an oversight mechanism.

No accuracy figure, model card, named models, hallucination rate or evaluation methodology was located. The closest thing to a performance statement is a customer testimonial saying they have not seen hallucination issues, which is impression rather than measurement. For a product generating differential diagnoses and dosing suggestions, published error characterisation matters more than it does for a note writer, and none exists.

Stronger than most in this category and one claim needs verifying before it is relied on. Avo states that randomised controlled trials have shown it saves clinicians around 50 percent of their time and improved emergency department admission delays by 30 percent. If accurate that would be among the best evidence in this index, but the claim appears on a marketplace listing with no citation, publication, author list or trial registration located in this pass, so it is graded as asserted rather than verified. Independent of that, named institutional customers are real and specific, including Children's of Minnesota, Driscoll Children's Hospital, PM Pediatric Care and Harbor Health, alongside implementation of a RAND consensus pathway for vertebral fragility fractures. Ask for the trial citations; they would move this to A.

Not assessed. No statement on audio or transcript retention, de identification or training use was located, which is a notable gap for a product that reads structured patient data including laboratory history, medications and clinical notes as well as capturing conversation.

Not assessed. No product specific business associate agreement posture was located in this pass.

Not assessed. No named or dated attestation and no trust centre located, which is notable for a vendor deployed inside children's hospitals and health systems where such reports are normally required at procurement.

No clearance claimed and none located, and this is the clearest software as a medical device scoping question raised anywhere in this wave. Avo generates differential diagnoses, treatment recommendations and dosing guidance embedded in the clinical workflow, which is decision support rather than documentation. Guideline grounding may well place it within the clinical decision support exemption criteria, since recommendations traceable to published sources that a clinician can independently review are treated differently from opaque model output, but that is precisely the analysis a buyer should ask the vendor to have documented rather than assume.

Graded up on a structural property rather than a disclosure document. Avo's recommendations are anchored to NAMED PUBLISHED SOURCES, including medical society guidelines, MCG recovery milestones and RAND consensus pathways, rather than to model inference or to an institution's own historical patterns. That is the same property that earned Nym Health credit in this index for being anchored to published AMA, CMS and WHO guidance rather than institutional intensity: a clinician can trace a recommendation to its source and disagree with it on the evidence. The no code builder also means the health system, not the vendor, decides which guidelines are deployed, which puts content governance with the organisation. Held at B because no fairness statement, subgroup analysis or accent and dialect performance disclosure was located despite support for 50 or more languages.

Specific and bidirectional rather than asserted, across four named systems. Avo is a listed MEDITECH vendor, available on the athenahealth Marketplace where it is enabled by the vendor and an athenahealth representative and appears in the left sidebar inside the patient chart, links to Epic Orders and Order Panels, and integrates with Juno EHR. It names the data it READS, including laboratory result history, vitals, medication history, clinical notes, diagnoses and demographics, which is unusually concrete. Held at B because integration certification levels and the depth of the Epic path specifically were not verified in this pass; confirmation there would move it to A.

Not assessed. Delivered inside the EHR and also usable outside it, but no hosting region, residency option or sub processor detail was located.

No published rate card. The vendor describes its pricing as competitive and unmatched in the AI scribe market for smaller hospitals and clinics, which is a positioning claim rather than a number, and a third party directory estimates a range of roughly 100 to 500 US dollars without stating the unit or tier. Graded on the house convention applied to Abridge: no public pricing is Not Rated, with the third party estimate recorded in the pricing record.

The broadest and most carefully specified coverage in this part of the index. Capture handles three or more participants across 50 or more languages, with the participant mix named explicitly as child, parent, translator and clinician. That is the encounter shape most ambient scribes quietly fail at, and it is the norm in paediatrics and in any consultation using an interpreter. The customer base matches, with several named children's hospitals. Clinician coverage extends beyond physicians to advanced practice providers, medical assistants, dieticians and behavioral health providers, and workflow coverage spans admission, rounding, discharge, charting and care planning rather than the outpatient visit alone.