Aidin

The artificial intelligence claim lives in the news library and nowhere in the product. A news post on Aidin's own site, summarizing Modern Healthcare's coverage of its chief executive, calls the company's platform AI powered and credits it with speeding up discharge. A second post carries the founder's remarks on artificial intelligence in care transitions.

Every page that describes the product says something else. The platform page lists a virtual print driver, a fax inbox, a document reader that sorts inbound faxes, prints and emails for assignment to patients and requests, secure messaging, assignment alerts for incomplete tasks and overdue work, patient checklists, reviews, and zip code level supply and demand data. The plans page repeats the same list tier by tier. The Epic page describes data flowing between the record system and Aidin. No model, method, training data, accuracy figure or learned behavior appears anywhere, and nothing says the document reader or the at risk flags are anything but rules and staff work.

Remove the claim and the product is unchanged: a two sided referral network, a directory, a document and communication hub, and a workflow tracker.

The division of labor is published and one part of it is not.

Nothing in this product decides anything. Case managers manage referrals, assign tasks and work deadlines; the platform surfaces visual flags for incomplete tasks, at risk flows and overdue assignments, and a person acts on them. Authorization tracking follows requests rather than submitting or deciding them. The choice of post acute provider is presented to the patient with quality, satisfaction and outcome data rather than made by the system, which puts the consequential decision with the person affected by it and matches what discharge planning rules require.

The terms close one gap that would otherwise sit here: core referral matching is free and equally available to every customer regardless of paid plan or integration, and fees do not vary with the volume or value of referrals. So paid standing does not buy the match itself.

What is unpublished is presentation. Whoever sets which providers appear, in what order and on which measures shapes a choice that is nominally free, and nothing describes the ordering, how many options a patient sees, how the hospital's preferred partners are treated, or how the measures are weighted. Nothing describes what happens when no provider accepts a referral, or when a deadline passes.

The product is described in detail and the technology inside it is not described at all.

Functional disclosure is good. The platform page enumerates the printer driver, the fax inbox, the document reader, secure messaging, assignment alerts, patient checklists, reviews and market data, and the plans page maps each capability to a tier, so what each tier includes is clear.

Beneath that there is nothing. No architecture, no method, no named component and no accuracy, precision or error figure appears for anything. The document reader, which sorts inbound faxes, prints and emails for assignment to patients and requests, publishes no accuracy rate, and a misassigned document puts one patient's clinical record in another patient's referral packet. At risk flags are not explained, so whether they are deadline rules or something learned is unknown, and those are different products. The quality analytics that shape a patient's choice of provider are the least specified: which measures, from which sources, over what period, combined how, is unpublished.

The gap matters more because the company's news library calls the platform artificial intelligence powered while no product page carries a technical claim of any kind.

The architecture is described, one third party is named, and the rest of the chain is not.

What is visible: a hosted platform with a virtual print driver at the customer, fax lines inbound and outbound, email ingestion, a document reader that sorts what arrives, messaging, and a provider directory with quality and review data. Jumio is named as the identity verification service used from the Verified tier upward, so a buyer knows at least one outside party touches user identity documents. The business associate agreement requires any subcontractor handling protected health information to be bound to the Security Rule and to substantially similar terms, and forbids any of it leaving the United States.

What is missing is the list. No cloud or hosting provider, fax transport, document recognition service or analytics supplier is named, and no subprocessor register is published, so a hospital cannot tell which parties handle its patients' records. Document recognition is the likeliest outside dependency, since turning faxes and scans into assignable records at scale is rarely built in house.

The quality data has a supply chain of its own that is equally undisclosed: provider quality, satisfaction and outcome measures originate somewhere, most plausibly federal quality reporting and survey instruments, and nothing states the sources, licensing or refresh cadence.

One named deployment now carries a dated figure with its method attached. AdventHealth, a nonprofit system of more than 50 hospitals across nine states, reports projected annualised savings of $71 million after adopting Aidin, alongside a reduction in length of stay for patients going to post acute care and faster placement turnaround. The company states how the figure was built: cost avoidance measured over four months following a five month ramp, with average length of stay compared between July to October 2024 and July to October 2025. The projection is the vendor's own and the period is short, but the comparator, the window and the ramp are all stated, which is enough for a buyer to test it.

The rest is customer names and quotations. St. Luke's University Health Network, Edward-Elmhurst Health, Vanderbilt University Medical Center, ProHealth Care and VHC appear in stories and endorsements that describe the problem brought to Aidin rather than the result. No peer reviewed study or independent evaluation appears, and no readmission figure is published for any site, which is notable because readmission is the measure the product is marketed on.

Two categorical commitments are published, and the schedule behind them is not.

The business associate agreement, linked from the terms and binding on every user, says Aidin may not de identify protected health information, and may not use or disclose de identified data derived from it, without the customer's prior written consent. That is the commitment that matters most, because de identified data is the usual route by which platform content is reused, and here it is closed by default. The same agreement forbids access to or transmission of customer protected health information outside the United States, and requires workforce privacy and security training within thirty days of hire with annual refreshers.

What is missing is duration and specificity. No retention schedule is published for referral packets, documents arriving through the fax and print paths, or messages, and no deletion process is described.

The workflow question remains open in the same place it was. An open market referral sends patient information to several competing providers so they can respond, and nothing published says how many receive a given packet, what each sees before accepting, or what a declining provider keeps. That last gap weighs most, since a declined referral leaves clinical information at an organization with no treatment relationship.

Status, agreement, tier and subcontractor terms are all published, and the agreement itself is on the site.

Business associate status is stated: the privacy policy says information provided through the platform may be protected health information and that Aidin's processing of it on behalf of healthcare customers is governed by its business associate agreement. The agreement is published in full at a linked page, so it can be read before contracting rather than requested.

The tier is clear, which matters for a two sided network. The terms of service say access to the service requires every user to agree to Aidin's standard business associate agreement, and that the agreement controls over any inconsistent provision of the terms. That answers the question this product's structure raises, because it means post acute providers on the free tier are bound by the same instrument as an enterprise hospital customer.

Subcontractors are addressed: where Aidin uses subcontractors that handle protected health information, it must bind each to the Security Rule and to substantially similar restrictions. The agreement also forbids access or transmission of customer protected health information outside the United States, requires breach reporting within five days, and forbids de identifying protected health information or using de identified data derived from it without the customer's written consent.

What a post acute provider that declines a referral may retain is governed by the agreement in general terms rather than for this workflow.

Obligations are published; independent evidence is not. The business associate agreement commits Aidin to safeguards appropriate to its size and operations under the Security Rule, to reporting breaches within five days and security incidents on a stated procedure, to binding subcontractors to the same requirements, and to privacy and security training for every workforce member within thirty days of hire with annual refreshers. The Enterprise plan is marketed as enterprise security, and messaging is described as health privacy compliant. All of that is the company's own word.

Nothing independent stands behind it. No attestation such as SOC 2, HITRUST or ISO 27001 is named, there is no trust center, no penetration testing statement, no vulnerability disclosure policy and no offer of security documentation under agreement.

The exposure is wider than one organization. This platform sits between hospitals and a national network of post acute providers, holding referral packets that combine clinical, functional and insurance information at a vulnerable transition, so a compromise reaches many organizations at once. The participant mix sharpens it: large health systems run mature security programs, while small skilled nursing and home health agencies on a free tier often do not, and both hold accounts on the same platform. Identity verification through Jumio applies from the Verified tier, not the free one. Nothing says how free tier accounts are scoped and reviewed or how the print component is secured.

The regime that governs this product is named in the company's own material and built against. Device regulation is not it: referral routing, document management and authorization tracking make no clinical claim, diagnose nothing and recommend no treatment, so no clearance is required and none is claimed.

What applies is federal discharge planning law. The plans page sells IMPACT Act compliance as an Enterprise feature, described as holding partners accountable for outcomes and keeping patient choice practices and choice lists compliant, and the open market design, which presents quality, satisfaction and outcome data and lets the patient choose, is the shape those requirements ask for. The terms add a second regime the company addresses directly: it states that core referral matching is free and equally available regardless of paid plan, and that fees do not vary with the volume or value of referrals, with the stated intent of complying with the federal Anti-Kickback Statute and its safe harbors.

What is missing is evidence rather than alignment. No attestation, audit or outside assessment of discharge list compliance appears, and nothing describes where the quality data shown to patients comes from, how it is refreshed or how it is validated, which is what a regulator examining the requirement would test.

Nothing published describes how the platform's ranking and presentation are governed or tested: no methodology, no fairness review, no subgroup analysis, no outside audit.

The exposure here runs through institutions and reaches patients through them. The platform ranks and presents post acute providers using quality, satisfaction and outcome data, and provider quality measures track the population a facility serves. Facilities caring for poorer, sicker and more socially complex patients score worse on readmission and outcome measures for reasons largely attributable to case mix rather than care, so a presentation that does not adjust for that steers patients away from safety net providers and toward those serving easier populations. Which measures are used, whether any risk adjustment is applied, how they are weighted and how recent they are is unpublished.

One related question is now answered and the other is not. The terms state that core referral matching is free and equally available regardless of paid plan and that fees do not vary with referral volume or value, which addresses the match. They do not say whether paid standing, profile completeness or the badges the company awards for responsiveness affect ordering or prominence in the list a patient sees.

Mechanisms exist to catch a mistake, and nothing stands behind any output.

What exists: documents are assigned to patients and requests by staff rather than filed automatically, the platform keeps documents searchable and auditable by patient, referrals and authorizations are tracked with visible deadlines and at risk flags, and providers on paid tiers can see the reviews, satisfaction scores and readmission data held on their profile. A person can therefore find and correct a misfiled document or a stalled referral.

What does not exist is any commitment. The terms disclaim medical advice, state that the service is not a recommendation, warranty or guarantee that any provider found through it is suitably licensed, qualified or able to meet a patient's needs, and say the company assumes no responsibility for consequences of action or inaction based on it. No uptime commitment, service level, accuracy warranty, indemnity or remediation position appears for either side of the network, and the business associate agreement's breach reporting is the only published obligation that runs anywhere.

Availability is the omission that bites hardest. Discharge planning runs to a daily clock and beds are held while placements are arranged, so an outage during business hours stalls transitions at every customer at once.

Named record systems, described as two way, with the detail concentrated on one of them.

The Enterprise plan sells integration with Epic, Cerner and MEDITECH, populating patient census, payer information, demographics, partner utilization and patient choice documentation automatically, plus single sign on that opens Aidin from inside the record with no separate account. The Epic page goes further: referrals are initiated from within Epic, information flows from Epic into Aidin, and notes, the after visit summary and media update automatically, with the company describing more than a decade of supporting Epic users.

The other half of the network takes a different route. It ingests from fax, email and a virtual print driver installed at the customer, which is interoperability with the actual state of post acute care, where skilled nursing, home health and rehabilitation providers frequently have no interoperable record system at all.

Verification and scope are the gaps. The integration claims are the vendor's own, with no marketplace or validated listing cited, and integration sits behind the top tier rather than being available at lower ones. Nothing states whether an acceptance posts back to the chart as a discrete event, and the API is named only as an Enterprise add on.

Residency is committed in writing and the rest of the deployment is not described.

The business associate agreement, which every user must accept, states that Aidin will neither provide access to nor transmit customer protected health information to any person or entity outside the borders and jurisdiction of the United States, including to that entity's own employees or agents. That is a contractual residency commitment rather than a marketing line, and it also closes the offshore support question that a buyer would otherwise have to ask.

The shape of the deployment is a hosted web platform reached by hospitals and post acute providers, with one component installed at the customer: a virtual print driver that sends printed documents into the platform. No hosting provider is named, no region within the United States is given, no tenancy or segregation model is described, and no backup, continuity or recovery position appears.

Segregation carries particular weight here. Competing hospitals and competing post acute providers are tenants of the same platform, and referral activity is commercially sensitive between them: which facilities accept which patients, how fast and on what terms. Whether the market insight product is derived from that activity is unaddressed in either direction.

The commercial model is disclosed in structure, with one published price and no tier figures.

The plans page sets out four levels on the post acute side. Free Access takes a fax number or an email address, needs no registration, and receives referrals from any Aidin partner. Verified adds identity verification through Jumio, team access for up to fifteen users, patient reviews and weekly intake reports. Community adds outbound referrals, three printer licenses, a digital fax dashboard and 400 outbound fax pages a month, with further pages charged at $0.11. Enterprise adds bidirectional record system integration, single sign on, IMPACT Act compliance tooling and a command center. Naming what sits behind each level lets a provider work out which one it needs before contacting sales, and it shows that record system integration is a premium feature rather than a baseline one.

What is absent is every figure above the free tier and the whole hospital side. No price appears for Verified, Community or Enterprise, no unit of charge is stated for hospitals, and nothing says whether charging follows the case manager, the bed, the referral or the facility. The terms do say fees do not vary with the volume or value of referrals, and that fees may rise annually on sixty days' notice.

Both sides of the care transition are named, and part of the span has evidence behind it.

On the acute side the buyers are hospitals, health systems and the case managers inside them, with dedicated pages for each. AdventHealth puts the product in a multi state system of more than 50 hospitals, and St. Luke's University Health Network, Edward-Elmhurst Health, Vanderbilt University Medical Center, ProHealth Care and VHC appear in stories and endorsements spanning academic, community and regional systems. That is validation behind the acute half, and it is what lifts this above a claim.

On the receiving side, post acute providers join a national directory, free of charge, through a fax number or an email address, which is how the network reaches organizations that have no record system at all.

What is still unstated is composition. Post acute care is not one setting: skilled nursing, inpatient rehabilitation, long term acute care, home health and hospice differ in eligibility, authorization and clinical criteria, and nothing says which the workflow supports best or how directory density varies by provider type or region. Density by market is what decides whether the open market produces real choice or a short list. Nothing addresses pediatric or behavioral health placement, both of which have scarce and distinct networks.