Apricot Health

Generative AI drafting the documentation is the product. There is no electronic medical record, no services business and no prior software underneath it, and the company describes itself as AI native. What it automates is the drafting of OASIS and visit documentation from captured content, which is the model doing the work rather than a workflow layer presenting someone else's output.

The oversight advantage here is architectural rather than procedural, which is the strongest kind. Because documentation is drafted from a guided interview the clinician completes after the visit, plus structured capture during it, the clinician supplies the content rather than a model inferring it from overheard conversation. That materially narrows the fabrication surface compared with ambient generation, and the company argues the point explicitly, positioning against approaches it characterises as riskier.

An in home visit guide reduces omission during capture, and the company's own representatives state plainly that the product drafts OASIS forms from patient documents but does not replace clinical decisions, with agency review keeping final decisions with the team.

Held at B because none of it is quantified: no accuracy or error rate, no confidence signal, no abstention behaviour when the interview leaves a required field unsupported, and no statement of what the agency review layer actually catches.

The company names transparent reasoning as a product property, meaning a clinician can see why a recommendation or draft entry was produced rather than receiving it unexplained. That is a real feature claim and more than several competitors offer. Nothing else is disclosed: no model or model family named, no accuracy figure for OASIS field completion or narrative generation, no evaluation methodology, no error taxonomy and no model card. For a product whose output is a regulated assessment instrument, per field accuracy is the disclosure that matters and none exists publicly.

Nothing identifies any party in the chain: no model or model family, no foundation model provider, no hosting arrangement and no sub processor list was located in two passes, and no training position was found in either direction. The content types are what a buyer should press on, because they are unusual and unlikely to share one retention treatment. Inside the home the product captures signatures, vital signs, photographs and voice notes, and uploaded documents feed the same pipeline.

Photographs need their own answer for the reason that applies across home based products: an image is identifiable in ways text is not, may capture the patient's body, their living conditions and other people present, and persists as an image that cannot be redacted the way a transcript can. Captured consents are a second category worth separating, since a signed consent is both evidence and personal data.

One question is specific to this workflow and is easy to overlook: what happens to captured content when a visit is abandoned or a referral does not proceed, since the material exists but no episode of care was created to govern it. Ask for retention per content type, the position on abandoned visits, and a sub processor list.

Strong commercial signals, no measured customer outcome. A Series A led exclusively by Insight Partners in October 2025, that firm's first investment in an Oklahoma company, is meaningful external diligence, and a co branded integration page on the Netsmart marketplace is a documented partnership rather than an assertion.

Against that, no customer is named anywhere, described only as ranging from regional operators to national leaders, and no case study, controlled comparison or independent evaluation exists.

The headline figure needs reading carefully. Start of Care documentation falling from 120 minutes to 15 to 30 minutes is footnoted as being measured against leading EMR time targets rather than against observed time at a customer. A target is not a baseline, and comparing measured performance to somebody's target overstates the delta by whatever gap already existed between target and reality. Ask for before and after timings at a named agency.

No retention period, training use statement or de identification posture was located.

The architecture is now clearer and it is materially different from most of this category, which changes the shape of the exposure rather than removing it. This is not primarily ambient capture of a patient conversation. The core flow is a guided interview completed with the clinician after the visit, whose answers, together with uploaded documents, photographs and other in home data, are transformed into structured documentation. So the principal generative input is the nurse's own account rather than a recording of the patient speaking.

That is a genuinely lower exposure design than continuous encounter recording, and it should be credited. A clinician answering structured questions produces far less incidental content than an open microphone in a patient's living room.

What remains is still substantial and specific. The product captures signatures, vital signs, photographs and voice notes inside the home. Photographs are the item that most needs its own answer, for the reason that applies across home based products: a wound photograph is identifiable in ways text is not, may capture the patient's body, their living conditions and other people present, and persists as an image that cannot be redacted the way a transcript can. Consents captured in the home are a second category worth asking about separately.

So the questions are narrower than for an ambient product but no less pointed. Establish the retention period for photographs specifically, the retention for voice notes, whether either is used to improve models, and what happens to captured content if a visit is abandoned or a referral does not proceed.

No health privacy compliance statement and no business associate agreement terms were located on any retrieved surface.

Agreements plainly exist. The product is deployed across named home health agencies including regional and national operators, and it is listed as a partner in a major post acute record vendor's marketplace, so every one of those relationships required one. The gap this axis measures is different: nothing is published, so a prospective buyer cannot establish terms, scope or subprocessor position before entering a sales conversation, and no one comparing suppliers can either.

One feature of this vendor's distribution deserves a specific question rather than a general one. Where the product is reached through a record vendor's marketplace, a buyer should establish which agreement actually governs. Marketplace distribution frequently leaves customers assuming their existing record system agreement covers the partner application, and it does not follow. Establish whether the business associate agreement is with this vendor directly, whether the record vendor is a party, and who is accountable in a breach originating with either.

A second question follows from the capture surface. The product collects consents in the patient's home. Where a vendor's software is the instrument through which consent is obtained, a buyer should establish who holds that record, how long it persists, and whether it is retrievable independently of the note it accompanied.

The attestation the vendor does hold is relevant context but is not a substitute. A control examination describes how systems are run; it says nothing about the contractual allocation of responsibility for protected health information.

Ask for the agreement, the entity named in it, and the subprocessor list.

The earlier assessment found no attestation. That is overturned. The vendor announced SOC 2 Type I compliance, which is a genuine independent examination and puts this record above the many in this category that offer only compliance language.

The report type is the substance here and deserves stating plainly, because it is the distinction most often blurred in this segment. A Type I report assesses whether controls are suitably designed at a single point in time. A Type II report tests whether those controls actually operated effectively across a defined period, usually several months. Type I says the design was sound on the day it was examined. Type II says the design held up in practice. They are different assurances and the gap between them is where most real control failures live.

The vendor states the type accurately rather than saying only SOC 2, and that honesty is worth crediting. Several vendors in this lane assert SOC 2 without ever naming a type, which leaves a reader to assume the stronger one. This one did not.

So the grade records a real attestation with a clearly stated limitation. A buyer should ask when the Type II examination is scheduled, since Type I is normally the first step toward it, and should treat the current report as evidence about design rather than about operation.

Also unresolved: no trust centre carrying the report, its date or a scope statement, and no ISO 27001 or HITRUST alongside it. Scope matters for a platform capturing photographs, signatures, vital signs and voice notes inside patients' homes.

No FDA clearance or device authorisation was located and none is expected, since the product completes a documentation and assessment instrument rather than making a diagnosis. The regulatory exposure that does matter sits in the CMS conditions of participation for home health and in the accuracy of the OASIS assessment itself, which drives both quality reporting and payment. Ask how the vendor supports an agency through an OASIS accuracy audit, since that is the regulatory event this product's output would be examined in.

The grade describes incentive structure and disclosure, and the counterweights are substantial and belong first. The company states that clinician informed recommendations keep coding and documentation aligned while transparent reasoning and agency review leave every final decision with the team, its representatives state directly that it does not replace clinical decisions, the founder ran a home health agency for seven years, and the design is described as informed by extensive field research with nurses.

The founder has also publicly convened discussion on how agencies should vet AI vendors before letting them touch clinical documentation, which is an unusual thing for a vendor to organise. Against that, OASIS is not a neutral form: it drives case mix under the Patient Driven Groupings Model, so a product that shapes OASIS responses is operating on the instrument that determines what a thirty day episode pays. That places it on the same payment unit as Andy AI, and it is the reason this sits at C rather than higher. No fairness, subgroup or demographic performance disclosure was located.

One mechanism carries this grade and one absence keeps it here. The vendor names transparent reasoning as a product property, so a clinician can see why a recommendation or a draft entry was produced rather than receiving it unexplained. That is a real correction mechanism at the point of review, because a clinician who can see the basis for an entry can reject it on that basis rather than on instinct.

The design also deserves credit for a lower exposure architecture than most of this category, since the principal generative input is a guided interview completed with the clinician after the visit rather than a continuous recording of the patient, and a clinician answering structured questions produces far less incidental content than an open microphone in someone's living room. The absence is the one that matters most for what this product outputs.

Its deliverable is a regulated assessment instrument whose fields drive payment grouping and quality reporting, and no per field accuracy figure, error taxonomy or evaluation methodology exists publicly. For a product of this kind, per field accuracy is the disclosure a buyer needs, because errors are not evenly distributed across an instrument and the consequential fields are a small subset. No warranty, indemnity or remediation commitment was located. Ask for accuracy by field group, and for what the vendor commits to when a scored field is wrong.

One integration is properly evidenced rather than merely claimed, which is worth more than a long unnamed list. Apricot integrates with Netsmart myUnity, the post acute care record, and appears on the Netsmart marketplace with a co branded page authored jointly with that vendor.

Being carried in a post acute EMR's own catalogue is a form of external validation, since the platform vendor applies its own review before listing, and it also tells a buyer the integration is supported rather than improvised. The product is separately demonstrated for Axxess users.

Held at B because Netsmart is the only integration described in any depth, the remaining post acute record systems that dominate this segment are not addressed, and no integration mechanism or standard is described.

No hosting model, cloud provider, region or residency commitment was located, and no subprocessor list or statement about third party model services was found.

The question flagged in the earlier assessment remains the specific one for this product rather than a generic gap, and the second pass confirms why. Capture happens inside patients' homes, where connectivity is frequently poor and often absent, and the product collects signatures, vital signs, photographs and voice notes there before any documentation is generated. So content exists on a device, in a house, for some period before it reaches anything the vendor controls.

Nothing published states what that period is or how it is protected. Establish whether capture works offline, what is held on the device until it syncs, whether it is encrypted at rest there, what happens if a device is lost between the visit and the sync, and whether an agency can remotely revoke or wipe. Those are operational and security questions at once, and for a workforce driving between homes all day they are not edge cases.

The generation step raises the ordinary questions on top. The guided interview and the documents, photographs and notes accompanying it are processed somewhere, and nothing establishes where, by whom, or what a model provider retains.

One distribution point belongs here. The product reaches customers partly through a post acute record vendor's marketplace and integrates with that vendor's platform, so a buyer should establish which organisation's infrastructure holds what at each stage, rather than assuming a single boundary.

Ask for the hosting region, the subprocessor list, the model provider, and the on device retention and offline behaviour.

No price, tier or pricing mechanism was located on any retrieved surface.

Worth checking the Netsmart marketplace listing during evaluation, since partner catalogues occasionally carry pricing or packaging detail a vendor's own site omits.

Narrow by design and deep within the narrow part. The product addresses home health and post acute care, and within that it began with the Start of Care visit, the most complex and time consuming documentation event in the segment, covering it end to end across medication reconciliation, wound documentation, calendar plotting, interventions and goals and narratives. That is instrument level work rather than generic summarisation, since OASIS is a specific regulated assessment.

Support is described as extending to other OASIS, therapy and routine visits. Graded B rather than A because the breadth beyond Start of Care is recent or announced rather than demonstrated, and because coverage does not extend past this one care setting.