Machinify

The grade describes the mechanism, not the quality, and this record sits close to a boundary this index has drawn before.

The artificial intelligence is real: the company acquired and then took the name of a software business built around a cloud data intelligence platform, and it markets purpose built models across the payment continuum. But three of the four constituent businesses were payment integrity operations, and the capabilities listed, subrogation, coordination of benefits and complex claim review, are historically labour intensive recovery work supported by technology rather than performed by it. The asset is also substantially the claims data and the plan relationships, which is the moat is the dataset case.

The boundary matters and is recorded openly. This index rejected Cotiviti from the same segment on the ground that its own framing was that technology serves expertise. Machinify is distinguishable because the surviving brand and self description are the software platform rather than the services, which is why it is graded rather than rejected. If Mike reads the two as the same shape, they should be treated the same way, and that is his call rather than one to make quietly.

Nothing published describes where automation ends and human review begins. Payment integrity work produces findings that a claims examiner, a nurse reviewer or a recovery specialist acts on, and the constituent businesses employ substantial numbers of such people, so a human is clearly in the process. What is absent is any statement of which determinations the models make alone.

That gap is more consequential on the payer side than the provider side. A payment integrity finding can result in a claim being denied, adjusted or recovered after payment, and the party affected is a provider or a member who was not party to the decision and may not see the reasoning. A published account of what is automated, what is reviewed and what can be appealed would be the natural disclosure and none was located.

No model, architecture, method or performance figure is published. The description available is a platform for the secure and transparent deployment of purpose built artificial intelligence, which names an intention rather than a mechanism.

The word transparent appearing in the company's own description of its platform without any accompanying disclosure is worth noting rather than passing over. Nothing published states how a finding is explained to the plan, to the provider whose claim is affected, or to a regulator asking how a determination was reached.

Nothing identifies any party in the chain: no model or model family, no hosting arrangement and no sub processor list was located in two passes, and no published retention schedule, encryption detail or model training position was found. The data concentration is the point worth naming and it is a stewardship question in its own right rather than a consequence of any control failure.

Combining four payment integrity businesses serving many of the largest health plans assembles claims data across a very large share of the insured population inside one organisation, and the company describes that data asset as a source of advantage, which is candid and is also the reason to ask.

Claims data at that scale is not merely a large version of a small holding: it supports inference about utilisation patterns, provider behaviour and populations that no single plan's data would permit, and the parties whose care it records have no relationship with the company and no way to learn it holds anything about them.

Concentration achieved through acquisition also means the underlying records arrived under terms agreed with four different sets of customers, and nothing states whether those terms were harmonised, whether data from the constituent businesses is combined, or what governs the combined asset. Ask whether the four data estates are joined, what governs the combined set, retention, and whether claims data trains models.

Commercial scale is substantial and independently reported: revenue above 500 million dollars, a reported valuation near 5 billion, and service to more than 60 health plans including many of the twenty largest, with some materials citing 75 to 85.

This index does not accept deployment scale as evidence of benefit, and here there is a further reason for caution. An independent analyst assessment recorded alongside those figures states that Cotiviti and Optum remain the entrenched leaders in enterprise payment integrity and that this company's market leadership is still emerging, and separately notes the integration burden of combining four businesses. That is a useful outside counterweight to a roll up narrative and it is recorded rather than omitted. No savings study, accuracy measure or published outcome was located.

Graded on an honest basis. No published retention schedule, encryption detail or model training position was located in this pass.

The data concentration is the point worth naming. Combining four payment integrity businesses serving many of the largest health plans assembles claims data across a very large share of the insured population in one organisation, and the company describes that data asset as a source of advantage. Concentration of that kind is a stewardship question in its own right, separate from any control failure, and it deserves an answer from the company rather than an inference here.

Graded on an honest basis and flagged for re verification. No compliance statement or agreement posture was located in this pass.

The structure is more complex than a single vendor relationship. A payment integrity organisation acts for the plan, handles member and provider data, and in subrogation work pursues recoveries from third parties entirely outside the health system. Expansion into government programmes adds federal contracting obligations on top. Which instrument governs which flow is not described publicly.

Recorded honestly: the dedicated trust and security search this index requires was not run in this pass, so the grade is provisional and should not be quoted until it has been.

An organisation contracting with many of the largest health plans and with government programmes will have passed extensive security assessment, so the absence here reflects what was retrieved rather than what exists.

No device pathway applies and none is claimed. Payment integrity is administrative and financial rather than clinical.

The regulatory surface that does apply is large and worth naming because buyers assess this segment against the wrong framework. Claims determinations sit under Centers for Medicare and Medicaid Services payment rules, state insurance regulation, and prompt payment statutes that constrain how long a plan may take and on what basis it may adjust. Government programme work adds federal contracting and audit obligations. Recovery after payment carries its own state law constraints. None of that is device law and all of it bears directly on what this product may do.

Nothing published on model evaluation, monitoring, error rates or appeal, and this is the axis where the omission matters most on this record.

The reason is regulatory as much as ethical. Several states have moved to restrict artificial intelligence driven coverage denials and to require that a qualified human makes or reviews the determination, a trend this index has already recorded on the payer side. An organisation applying models to claims at this scale is operating directly into that scrutiny. Nothing published states whether a model can cause a claim to be denied without human review, how a provider learns why, or what recourse exists. A payment integrity finding is not a clinical error a clinician catches; it is a financial determination the affected party may never see explained.

Two passes located no model, architecture, method or performance figure, no evaluation methodology and no warranty, indemnity or remediation commitment. The available description is a platform for the secure and transparent deployment of purpose built artificial intelligence, which names an intention rather than a mechanism.

The word transparent appearing in a company's own description of its platform without any accompanying disclosure is worth noting rather than passing over: transparency is not a property a vendor can assert about itself, it is a property a reader confirms by finding something to read, and here there is nothing. What that absence covers matters because of who bears the consequence.

This operates in payment integrity, so an output affects whether a provider is paid for care already delivered, and nothing published states how a finding is explained to the plan, to the provider whose claim is affected, or to a regulator asking how a determination was reached. The provider is the party with the loss and the least access: they receive an outcome rather than a reasoning, and no published appeal mechanism, error rate or reversal rate exists to work from.

A false positive here is a recovery demand against a provider who did nothing wrong, and it is absorbed as a cost of doing business rather than surfacing as a measurable vendor error. Ask for the false positive rate, the overturn rate on appeal, and what a provider is shown about the basis of a finding.

Largely not applicable in the usual sense and graded rather than penalised. The systems that matter here are claims platforms, enrolment and eligibility systems and pharmacy benefit data on the payer side, not the electronic health record.

The integration burden is nonetheless real and mostly internal: four merged businesses each brought their own data models and workflows, and independent commentary describes harmonising them as an ongoing programme rather than a completed one. For a buyer that is a question about which platform they are actually being sold.

Described only as a cloud based platform. No hosting region, retention schedule or customer controlled option was located.

A specific question follows from the roll up rather than from the technology: whether client data from the four constituent businesses now sits in one environment or remains separated, and whether a plan that contracted with one of the predecessor companies has consented to its data joining a combined asset. Nothing published addresses it.

Nothing published: no price, no mechanism, no unit of sale.

The mechanism is the thing to ask about here rather than the number, because payment integrity is frequently sold on contingency, meaning the vendor takes a percentage of what it recovers or prevents. That model aligns the vendor with finding more, which is efficient and also creates an incentive worth understanding on a product that decides whether claims are paid. A buyer should establish whether pricing is contingency, fixed fee or hybrid, and what happens commercially when a finding is successfully appealed.

Wide across the payer side and absent from the provider side. Capability spans subrogation, coordination of benefits, pharmacy payment integrity and complex claim review, which are four distinct disciplines rather than one product applied broadly, and each arrived with a different constituent business.

The buyer is a health plan, with stated reach to more than 60 and by some accounts 75 to 85, including many of the largest, plus expansion into government programmes where the payer is a public body. Coverage is specialty agnostic because claims arrive from everywhere. United States only, which is inherent since payment integrity is a function of a specific reimbursement system.