HealthHelp wants fewer denials. The interesting part is how.

A rejection has a sequel. Somewhere, a physician’s office gathers more paperwork. Someone checks a status screen. Someone else prepares an appeal. The original decision may have been inexpensive to make; the argument that follows has its own running costs.

An unnamed regional Blues plan offers a particularly revealing example. In HealthHelp’s account, the plan had high denial rates and dissatisfied providers. Its savings depended on saying no. HealthHelp introduced proactive outreach and specialist education. Providers could modify a request or withdraw it after discussion. The company reports that denials fell by half while savings increased.

A rejection has a sequel

The Blues case reports an annualized savings rate of $8 million, a net-direct-savings ROI of 4.9 and a provider Net Promoter Score of 60. These are HealthHelp’s published results for one client. Their interest lies in the mechanism: expenditure can change because a clinician chooses a different path, before a formal refusal is necessary.

There is a useful distinction here. A request withdrawn after discussion and a request denied outright may both disappear from an approval queue. They tell different stories about what happened. A buyer assessing this model should want to know whether the patient received a suitable alternative and whether the conversation shortened the journey. The denial count alone cannot answer either question.

The phone call inside the platform

HealthHelp occupies the space between an ordering clinician and the insurer paying for care. Its HealthHelp Platform combines clinical decision support, AI assistance, intake channels and analytics. The company offers specialty programs rather than a consumer subscription. A health plan buys the service; a medical office encounters the workflow.

Humana’s cardiology materials make the process concrete. An office supplies the patient and provider identifiers, the requested procedure, the diagnosis and supporting clinical information. Requests outside evidence-based criteria go first to nurse reviewers. When needed, the ordering physician can consult a board-certified cardiologist or interventional cardiologist.

An incomplete request can prompt clarification. A clinically complicated request can prompt consultation. Those are different problems, and treating them as different problems is the point. The software helps route the work. The specialist supplies judgment where a form cannot settle the matter.

Born in the imaging room

The present-day AI pitch has an older foundation. Cherrill Farnsworth founded HealthHelp in 1999. Its diagnostic imaging program began that year with an emphasis on selecting appropriate tests, protecting patients and reducing unnecessary radiation exposure. Before the current fascination with machine intelligence, there was already a question about whether a requested scan was the right scan.

Imaging remains part of the business. HealthHelp’s radiology program covers MRI, MRA, CT, PET and nuclear imaging. Its other programs reach into cardiology, oncology, musculoskeletal care, genetic testing, sleep therapy and surgical management. The common thread is a decision about a particular service, made with evidence and specialty knowledge.

Sleep care shows why that knowledge must be specific. Humana’s published materials describe a review that considers whether an in-laboratory study or a home test is appropriate, including relevant conditions and patient information. HealthHelp’s sleep program also targets duplicate testing. A general instruction to “spend less” would be a remarkably poor substitute for understanding which test belongs in which case.

Software with a clinical workforce attached

In March 2017, WNS announced that it was buying HealthHelp for $95 million, before adjustments for cash, debt and working capital. At the time, the announcement described approximately 400 US medical and operational professionals and a peer network of more than 100 doctors and 11 university medical systems. These are historical figures, but they reveal what the buyer valued: a clinical operation alongside a proprietary platform.

Today, HealthHelp sells technology and specialty-management services to payers, including fully delegated utilization-management programs. Its model therefore asks a customer to consider both the software and the people handling the exceptions. The relevant cost comparison includes the plan’s own review workload, provider interactions and medical spending, rather than simply the price of a software seat.

The market has substantial alternatives. EviCore by Evernorth and Carelon Medical Benefits Management offer overlapping specialty review services. Carelon also describes evidence-based guidance, clinical expertise and collaboration. HealthHelp’s positioning emphasizes proactive outreach, customization and independence from health-plan ownership. None of those claims makes clinical judgment its exclusive property.

Capgemini completed its acquisition of WNS in October 2025. HealthHelp now belongs to that broader business and technology services group. For a prospective payer, the ownership chain is context; the daily experience still turns on whether the request moves correctly through the clinical process.

The partners do different jobs

HealthHelp’s partnerships make more sense when separated by the problem each addresses. Covera Health adds radiology provider matching. Their November 2021 agreement connected Covera’s Centers of Excellence program with HealthHelp’s diagnostic imaging workflows. After deciding which test is appropriate, the system can help address where the patient receives it.

InformedDNA contributes genomics expertise and clinical criteria. The alliance extended HealthHelp into standalone genetic-testing utilization management and supported related cardiology and oncology services. It is an instructive choice: expanding the list of specialties requires knowledge specific to those specialties, not just another menu item in a portal.

Humata Health tackles the handoff between electronic medical records and authorization systems. The 2024 alliance describes FHIR-based integration intended to reduce manual entry and let providers manage authorization within the EMR environment. Availity supplies another route into established workflows. An April 2025 agreement announced integration of HealthHelp’s specialty outpatient guidelines into Availity’s Intelligent Utilization Management solution, alongside access to delegated programs and clinical experts.

The most interesting part of the platform is the conversation it makes room for.AN EDITORIAL OBSERVATION

A faster queue still needs a careful reader

The collaboration with Geisinger Health Plan and Anterior is HealthHelp’s clearest public example of its clinical AI approach. A company-published case study reports a 50-70% reduction in turnaround time for escalated cases requiring clinical review, and a 76% reduction in approvals requiring human review. The scope matters: those are specified workflow measures, not a promise that every authorization becomes 70% faster.

“We needed a way to improve both the speed and the accuracy of decisions,” the case study quotes Geisinger’s Medical Management Director as saying. That pairing is essential. Automating an ill-informed decision would merely let a mistake travel faster. HealthHelp’s responsible-AI paper argues for using automation to gather information and accelerate approvals, while bringing experts into complex cases.

For an operations team, the copyable idea is the division of work. Give routine cases a clear path. Give ambiguous cases access to a person who understands the subject. Count the rework as well as the completed transactions. This is an interpretation of the model, and it depends on sound criteria, useful documentation and specialists who can actually engage with the ordering provider.

The October handoff

Highmark’s molecular-testing transition puts those conditions into view. Its June 2026 notice set October 1 as the start date for HealthHelp to manage genetic and molecular lab authorizations. In Delaware, Pennsylvania and West Virginia, HealthHelp replaces EviCore; in New York, the notice describes expanded prior-authorization requirements.

Providers start in Availity Essentials and are routed through Predictal to HealthHelp. Highmark also warns that the initial transition may require re-entering member and provider information. That small disclosure deserves attention. Even a system built around collaboration can ask the doctor’s office to type the same thing twice.

HealthHelp’s model is most persuasive where specialist judgment can resolve uncertainty and integration can remove avoidable work. Missing documentation, disconnected systems or an unsuitable clinical pathway can erode that advantage. The practical lesson is to inspect the handoffs. The phone call, the record and the review must meet at the same case. Otherwise, the patient is still waiting while the software congratulates itself on being quick.