Affineon Health

Affineon sells an AI inbox, and its two modules do their headline work by protocol. Labs Core sorts results by settings each practice configures: selected normal results, and abnormal results inside parameters the provider sets as clinically insignificant, are closed out of the inbox and the patient is notified; CC'd results are routed to staff; result sets are held until complete; and ASCVD and CKD risk are calculated from ACC/AHA and ISN guidance.

Rx Renewals runs each request through a configurable protocol of guidelines (a visit within a set number of days, a lab panel within a set window, a controlled substance agreement on file) and routes it to staff when the criteria are met or to a provider when they are not. Those are functions rules can perform, and no learned method is described behind them.

The learned component the company does describe is language work: Labs Core 'speed-reads the entire patient chart' to summarize and trend labs, and for results that need the provider it drafts a proposed note to the patient, which the provider reviews and sends. The AI is a feature layer on a product whose inbox reduction rests largely on configured protocols.

The routing rule is published for both modules. Labs Core closes selected normal results, and abnormal results inside parameters the practice sets, without a provider seeing them, and notifies the patient; results that need the provider's attention are summarized and trended, with a drafted note to the patient that the provider reviews and sends.

Rx Renewals marks each guideline met or not met, sends requests that meet the protocol to staff and those that do not to a provider, and some practices route every scheduled medication to a provider by default. The thresholds are set by the practice ('All settings are configurable at the practice level so doctors stay in control'), and the Embed page sells the agent as triaging the provider inbox '24 hours a day, 365 days a year', which is the scope this oversight has to cover.

What is missing is what happens after the gate is wrong: nothing describes an audit or sampling of results closed without review, a result type that can never close automatically (critical values, a new abnormal trend), how a provider reopens an item the agent closed, or how often a result closed as insignificant later proved significant. That is the category's decisive failure, a routine looking result that was not routine, and on the closed side no clinician sees it. The line that every renewal request 'receives consistent, intelligent, and clinically grounded review' describes uniform handling, not a claim that every output is right.

The company describes the product as agentic AI built on 'Affineon's proprietary technology', says 'Our AI' processes normal results and drafts patient notes, and says the agent 'speed-reads the entire patient chart'. No model or model family is named, nothing says whether the language capability is built in house or licensed, no training data is characterized, and no versioning or release practice is published, so nothing marks when the system changed.

The protocol side is described more fully than the model: the Rx Renewals page shows the guideline types, the difference between an evaluator guideline that decides pass or fail and a display only one, and the markers a user sees.

The company names no party that sits between a patient record and an answer. The product reads lab results, chart data and renewal requests from the practice's record system (athenahealth is the one named) and writes back into its inbox, but no model provider, hosting provider or subprocessor is named, and no subprocessor list is published.

The only privacy policy (effective 14 December 2023) says the company may share information with 'third-party service providers who assist us' without naming any, and it addresses information visitors and account holders provide rather than patient data. The product's chart summaries and drafted patient notes come from its AI, yet nothing says who receives the chart.

The headline claim sets the bar: the home page says Affineon reduces inbox volume by 50 to 60 percent and saves providers and their care teams 3 to 5 hours per week, and its summary figures on the same page say 30 to 80 percent and 5 hours a week. The evidence behind it is vendor published.

The September 2025 case study names Annapolis Internal Medicine and Dr. Titus Abraham and reports that monthly inbox volume dropped 53 percent after adoption, with about 3,664 lab results handled automatically and 3,500 renewals expedited each month; it gives no baseline period, no measurement method and no provider count. Further case studies are listed for Dr. Painter and Dr. Kahney, and testimonials on the product pages report '45 minutes a day reviewing labs' saved (Dr.

Schroeder) and a 40 to 45 percent cut in laboratory administrative workload (a customer administrator). No peer reviewed study or outside evaluation is published, and nothing measures what matters most for a product like this: how often a result closed as normal or insignificant should have reached the provider, and how often drafted patient notes are sent unedited.

What is published is a general assurance: 'We're serious about security. Affineon is SOC2 and HIPAA compliant.' The only privacy policy (effective 14 December 2023) says it covers 'our website, products, and services', but its content addresses names, email addresses, phone numbers, IP addresses and cookies, marketing email, and an identity resolution service (Retention.com) that links website visits to other personal information; it says nothing about lab results, patient messages or chart data.

Nothing states how long lab results, chart data or drafted notes are retained, whether customer or patient data trains the AI, or whether anything is de-identified. Safety engineering, meaning guardrails on drafted patient notes, hallucination mitigation and how a safety event is handled, is not published either.

The site states 'Affineon is SOC2 and HIPAA compliant' beside a HIPAA Compliant badge. That is a compliance claim, not a statement of business associate status, and no business associate agreement, its terms or the tier it applies at is published.

The only privacy policy (effective 14 December 2023) names products and services in its opening sentence but addresses contact details, cookies, marketing email and website analytics, and contains nothing on protected health information, for a product that reads lab results and chart data and notifies patients.

An AICPA SOC badge on the site, beside the statement 'Affineon is SOC2 and HIPAA compliant', names a recognized audit of how the company operates security and presents it as the company's own rather than a host's. The report type (Type 1 or Type 2), its date and its scope are not stated, no trust center or report request route is published, and no other certification is named. SOC 2 produces an attestation report, so 'compliant' is loose wording for it.

No clearance is claimed, and the company states no regulatory position. The Labs Core page says the product 'can automatically triage the provider's inbox, handling normal and clinically insignificant lab results while notifying patients and closing them out of the inbox', and 'Our AI automatically processes selected normal lab results and instantly notifies the patient with the good news.'

What sits behind that sort is the practice's own configuration: providers select which normal results close automatically and 'customize parameters' for which abnormal results count as clinically insignificant, and 'Protocols and patient messages are customizable for every practice.' Results outside those settings go to the provider with a summary, trends and a drafted patient note the provider reviews and sends.

Rx Renewals checks each request against guidelines the practice configures, shows each guideline as met or not met, and sends the request to staff or to a provider. The tension is on the closed side: a result closed as normal or insignificant, and the message the patient receives about it, is never seen by a clinician.

How the AI's behavior is governed or tested is not published. The site, the Annapolis Internal Medicine case study and the privacy policy carry no responsible AI policy, no evaluation of how accurately results are classified as normal or insignificant, and no results by patient group, lab type or practice.

The company values ('Do the right thing') are general statements about the business, not about the model, and the ACC/AHA and ISN guidance cited for the risk calculations describes the formula used, not how the AI is tested. For a product that closes results without review, nothing published says how classification accuracy is measured, across which lab types and patient groups, or who reviews it.

Mechanisms exist that let someone challenge an output, and nothing stands behind the output. For renewals, each protocol shows every guideline as met or not met at the top of the action note, so the basis for sending a request to staff is visible, and requests that fail go to a provider. For lab results that reach the provider, the drafted patient note is a proposal the provider reviews before sending.

Nothing covers the results the product closes on its own: no accuracy or error rate is published for normal and insignificant classification, no warranty, indemnity or remediation commitment attaches to the output, and no route is described for a patient whose result was closed when it should not have been. The home page offer ('Try Affineon risk-free. Satisfaction guaranteed.') publishes no terms, remedy or test and concerns the purchase rather than the output.

Affineon works inside the practice's record system inbox rather than as a separate application, and athenahealth is the one system named: the home page shows the athenahealth logo under 'Integrated with', each product page cites a 4.8 rating on the athenahealth Marketplace, and the company announced joining athenahealth's Marketplace Program.

By its own description the product reads results, chart data and renewal requests and writes back, closing items out of the inbox and placing protocol results in the action note. No other record system is named, and the marketplace listing is cited by the vendor rather than quoted. The Embed program offers the agent to EHR vendors, AI scribes and other health technology companies to sell under their own brands (Commure has joined), which may be the route to other systems. Whether it runs on any other record system today is not stated.

Nothing is published about where the system runs or where the data rests. The company says the product works inside the EHR inbox with no new application or login, which describes where users see it, not where lab results, chart data and drafted notes are processed or stored. No hosting provider, region, tenancy model or on premise option is stated on the site or in the privacy policy.

Affineon publishes a full per provider per month price ladder on its home and product pages: Labs Core or Rx Renewals at $100 (1 to 7 providers), $95 (8 to 19), $80 (20 to 99) and $75 (100 to 249), the bundle of both at $175, $165, $155 and $145 across the same bands, and 250 or more providers quoted as enterprise. The unit of charge is the provider per month, and the Labs Core page puts the cost at 'around $4 per day'.

There is no route to start without a sales conversation: every call to action is a demo request or 'Contact Sales', and the 'Try Affineon risk-free. Satisfaction guaranteed.' offer publishes no trial length, terms or refund mechanics. Tiers, figures and a unit are published; a self serve start is not. The Embed page offers white label partners the agent for 'less than $3 per day', with no unit or terms.

Scoped to ambulatory practice and to two workflows rather than to the whole inbox. Published pricing bands run from a single provider to 250 and above, so the commercial model addresses solo practice through enterprise, and the named customer evidence is primary care and internal medicine. Affineon states it has processed more than 2,000 distinct laboratory types, a substantive breadth claim within its scope. No inpatient, emergency or specialty specific coverage is described, and inbox items outside labs and renewals, including general patient portal messages and staff messages, are not claimed.