Optum Integrity One

The coding is model driven and a substantial share of the stated value is not. Clinical Language Intelligence reads the record and assigns codes autonomously at enterprise scale, which is the model doing work people previously did.

What holds this below the A that Fathom and Arintra earn is that the product is explicitly a consolidation play as well as an engine. Published benefits lead with reduced need for multiple vendors, streamlined operations and lower total cost of ownership, and the platform is described as unifying several separate revenue cycle functions into a single interface. A buyer replacing four vendors with one retains real value from the interface, the workflow and the reporting even if the models underperform, which is not true of a pure coding engine. Graded B on that distinction rather than on any weakness in the underlying technology.

The escalation concept is described and the boundary is never quantified.

What is published is a routing principle: both fully autonomous and partially autonomous coding are offered, and the platform brings a human expert in the loop when clinical signals indicate greater complexity. The company illustrates it with a worked emergency department example. That is a clearer articulation of intent than several competitors manage.

No number attaches to any of it. No automation rate, direct to bill share, accuracy figure or confidence threshold is published, so a buyer cannot tell what proportion of encounters the platform actually completes without a human, nor what triggers escalation beyond the phrase clinical signals. No service level agreement, validation trial or benchmark offer was located. Adherence to responsible and ethical artificial intelligence practices is asserted without content.

Every vendor graded B or above on this axis publishes at least one boundary number. This one publishes none, which is what places it at C. Ask for the autonomous completion rate by setting and the specific escalation criteria.

A named, patented proprietary technology and almost nothing underneath it. Clinical Language Intelligence is identified as the engine across both this platform and the older assistive coding products, and it carries patent protection, which is more than a marketing label. Processing is described as analysing the patient record and applying coding logic in real time as documentation is created, which is a meaningful architectural statement about timing.

Below that the disclosure stops. No model class, architecture, version, foundation model or training data provenance is described. Output scope is not enumerated for this platform: which code sets, whether professional and facility coding are both covered, and how evaluation and management levels are handled are all unstated, where competitors routinely list them.

No accuracy figure appears anywhere, in aggregate or by coding element. The published outcome figures measure productivity and documentation integrity financial impact rather than correctness. Graded C on the strength of a named and patented mechanism with no specification behind it.

Nothing in the chain is named. No foundation model provider, model class or version, no hosting arrangement or cloud platform, no sub processor list, and no statement on whether customer documentation contributes to model development. The patented language technology is proprietary, which indicates the language layer is built in house and says nothing about what sits beneath it.

One question specific to this vendor has no counterpart anywhere else in the index and is unanswered. The corporate group holds claims, clinical and pharmacy data at a scale exceeded by no other organisation in United States healthcare, spanning a large insurer, a pharmacy benefit manager and a care delivery arm serving tens of millions of people. Whether any of that data informs the training or tuning of the coding models sold to providers is not addressed in either direction. For a provider weighing whether its documentation patterns become visible to a corporate family that also adjudicates its claims, that is the first question rather than a technical footnote.

Ask what data the models are trained on, whether any of it originates elsewhere in the corporate group, and for the sub processor list.

Outcome figures exist and no customer is named for this platform.

Published results are a pilot at over 20 percent coding productivity improvement, and one large health system at 180 percent documentation integrity financial impact with a 50 percent rise in inpatient coding productivity. Both organisations are anonymous. The named references on the company's site, including a university faculty practice and two named hospitals, attach to the older assistive coding and documentation products rather than to this platform, and should not be read across to it.

Third party signal is a 2026 analyst assessment naming Optum a leader among revenue cycle management platforms. That is a company level ranking covering a much broader portfolio, it reflects analyst judgement rather than measurement, and it predates meaningful deployment history for a platform launched in May 2025.

No accuracy validation, peer reviewed publication, independently audited measurement or buyer survey score specific to this product was located. Graded C. For a platform from the largest vendor in this market, the absence of a single named reference site fifteen months after launch is the notable fact. Ask for reference customers running Integrity One specifically.

No stewardship material tied to this product was located. There is no retention schedule, encryption statement, access control description, data ownership or deletion position, data minimisation commitment, or statement on whether customer documentation contributes to model development.

As on the privacy posture axis, this is graded conservatively because the pass was product scoped and an organisation of this size certainly documents these controls somewhere. The finding is that none of it is surfaced where a buyer evaluating this platform would look.

The stewardship question that is genuinely distinctive here cannot be resolved by finding a corporate security page, because it is structural rather than procedural. Provider documentation processed by this platform sits inside a corporate group containing the largest health insurer in the country, a pharmacy benefit manager and a care delivery business. What separates the provider data handled here from the rest of that group, whether contractual, technical or organisational, is the question, and nothing published describes any such separation. Competitors answer a simpler version of this by committing that customer data never trains their models; there is no equivalent commitment here. Ask what firewall exists between provider documentation and the payer side of the group.

No health privacy material tied to this product was located in a dedicated pass: no compliance statement, control enumeration, de identification position, or business associate agreement posture, template or execution requirement.

This grade is deliberately conservative rather than punitive and the reasoning should be read with the record. An organisation of this scale, operating as a business associate to a large share of United States hospitals, unquestionably maintains extensive privacy infrastructure and executes agreements as routine practice. What was not found is any of it surfaced at the product level, and company level compliance material was not exhaustively searched in this pass. The axis should be revisited.

One piece of context a buyer will raise regardless of what the product page says. Change Healthcare, an Optum company, suffered a ransomware incident in February 2024 subsequently reported as the largest breach of protected health information in United States history. That event concerns a different entity and a different system, and it is nonetheless material to any assessment of protected data handling within this corporate group, and the product's published material does not address it.

No credential tied to this product was located in a dedicated pass. No controls report, information security certification, health security framework certification or cloud authorisation is presented on the product material, and no trust center, report availability process, penetration testing disclosure or vulnerability disclosure policy was found at product level.

The grade is conservative for the reason recorded on the privacy axes: an organisation operating at this scale certainly holds security credentials, and the pass was product scoped rather than an exhaustive search of a very large corporate estate. The finding is the absence of product level surfacing, not a conclusion that credentials do not exist, and this axis should be revisited.

Worth recording for the comparison a buyer will actually make. Vendors in this lane a small fraction of this size publish considerably more here: one publishes a daily updated live control report covering 244 monitored controls open to any prospective customer, another holds three named certifications with a trust center, and a third holds a state government cloud authorisation. Scale is not what determines this grade, and on current evidence the largest vendor in the market surfaces the least. Ask for the controls report, its type and period, and any health specific certification held.

No device pathway applies and none is claimed. Assigning billing codes and surfacing documentation opportunities are administrative determinations rather than clinical ones, so the absence of a clearance is correct.

The ordinary exposure is the usual one: codes on a claim are representations to a payer, and error is governed by federal false claims enforcement landing on the billing provider.

The extraordinary exposure is specific to this vendor's ownership and is not addressed anywhere in the product material. The platform surfaces documentation opportunities that increase captured clinical acuity, the company's worked example being the presence, type and acuity of heart failure routed to a documentation integrity team for a provider query. Acuity capture is the mechanism at the centre of sustained federal scrutiny of the corporate parent's own risk adjustment coding practices. A tool that improves acuity capture, sold to providers by that corporate family, occupies a position no competitor in this lane occupies, and nothing published describes the governance separating the two. Graded C because the regulatory position of the software itself is correctly represented while the ownership question is untouched.

A monitoring surface exists and no governance content sits behind it. The platform provides comprehensive reporting and dashboards with dynamic, scheduled and ad hoc real time data, described as delivering actionable insight for performance monitoring. That is an analytics capability rather than a governance instrument, and nothing states that it reports the distribution that would make drift visible.

Responsible and ethical use of artificial intelligence is asserted as a practice the platform adheres to. No content accompanies the assertion: no bias or fairness testing, no model validation methodology, no distribution of assigned codes against an expected benchmark, no breakdown by specialty, payer or physician, and no external audit of coded output.

The headline outcome makes the gap conspicuous. A reported 180 percent documentation integrity financial impact is a revenue capture figure, and revenue capture rising sharply is both the intended commercial effect and what a drift problem looks like from outside. Arintra and XpertDox both publish the billing level distribution reporting that separates the two; this platform publishes the financial impact without the monitoring. Graded C. Ask what evaluation and management and payment grouping distributions look like before and after deployment.

No performance figure is published, so there is no stated level against which a shortfall could be measured and nothing to hold the vendor to. No accuracy, precision, automation rate or confidence threshold appears anywhere in the product material.

No service level agreement, warranty, indemnity or remediation commitment was located, and no denial or reversal rate for codes the platform assigned is published. No validation trial, benchmark offer or pilot term of the kind RapidClaims, XpertDox and AccuCode publish was found, so a buyer has no published route to establishing performance on their own data before committing.

The allocation question carries an additional edge here. Codes generated autonomously by a platform owned by a major payer, and submitted by a provider to payers including that payer's own plans, create an apportionment problem no competitor presents. If an automated code is later challenged, the provider carries the false claims exposure while the tool that produced it originates inside the corporate family adjudicating the claim.

One pre emptive note: further productivity or financial impact figures cannot move this grade, since both measure commercial benefit rather than accuracy or recourse. Only a published accuracy measurement, a stated autonomy threshold, or a contractual commitment on either will change it.

Integration is implied throughout and specified nowhere. The platform is described as analysing the patient record as documentation is created and delivering insight directly to documentation integrity teams within the platform, which requires deep and timely access to the record system, and a dedicated pass located no statement about how that access works.

No record system is named, no interface standard is described, no connection mechanism is specified, no vendor marketplace or programme listing was found for this product, and nothing addresses whether coded output writes back to the billing system or is exported.

What can be established indirectly is that the parent organisation manages revenue cycle and information technology operations for health systems nationwide, so practical integration capability across major platforms is not in doubt. Inference from a parent company's service footprint is not product disclosure, and this record does not grade it as such. Graded C for credible capability with no published specification. Ask which record systems are supported, through what standards, and whether the real time documentation analysis requires an on premise component.

Nothing was located. No hosting model, cloud provider, region, tenancy model, residency commitment or customer controlled deployment option appears in the product material.

Graded at the same level as every other record in this lane that is silent on the question, which is the consistent treatment rather than a judgement that this vendor is better or worse placed than those peers.

Two factors would make a specific answer more valuable here than elsewhere. Real time analysis of documentation as it is created implies latency sensitive processing, which constrains where compute can sit and therefore says something about the deployment that is not stated. And the corporate group's scale means a shared multi tenant environment would sit alongside data from an enormous number of other healthcare organisations, including entities on the payer side, which raises a segregation question a smaller vendor does not present. Ask for the hosting arrangement, the tenancy model, and what isolates one health system's documentation from the rest of the group's estate.

Nothing about cost is published. A dedicated pass located no pricing page, no unit of charge, no range, no implementation or onboarding fee position, no minimum commitment, no pilot or trial terms, no return calculator, and no percentage saving against existing coding spend.

Cost benefit appears only as adjectives. Lower total cost of ownership, cost effective workflow and reduced need for multiple vendors are stated as benefits with no figure attached to any of them, which is weaker than the directional percentages that lift several competitors above D.

The consolidation positioning makes the omission more consequential than usual. A platform whose central economic argument is replacing several vendors with one invites a total cost comparison, and none of the inputs to that comparison is published. Ask for the pricing mechanism, whether modules are licensed separately, what the migration cost from incumbent point solutions looks like, and how pricing relates to the outsourced coding services sold by the same organisation.

The setting is hospitals and health systems and the detail stops there. Published material centres on the middle revenue cycle from point of care through final coding, with an emergency department worked example and an inpatient coding productivity figure, so acute care is clearly in scope.

What is not published is the boundary. No specialty enumeration exists, no statement covers whether professional and facility coding are both handled, no ambulatory or outpatient position is given for this platform specifically, and nothing addresses which settings are supported autonomously versus assistively. The older assistive products in the same portfolio are split explicitly between professional and enterprise editions, and whether Integrity One spans both is unstated.

Buyer size is implied as enterprise given the consolidation argument and the reference to coding at enterprise scale, which suggests this is not aimed at the physician group segment several competitors target. Nothing addresses coding regimes outside the United States. Graded C for a clear setting with no enumerated coverage inside it.