Medicare’s WISeR model is not just another health policy pilot. It is a government-backed distribution channel for AI companies, and the backlash will decide how far that channel can go.
For years, AI health startups have promised to make insurance approvals faster, cleaner and less painful. Now CMS is giving a small group of them something much more valuable than a demo environment: direct responsibility inside Original Medicare.
The Wasteful and Inappropriate Service Reduction model, better known as WISeR, began on January 1, 2026 and is scheduled to run through December 31, 2031. It covers selected services in Arizona, New Jersey, Ohio, Oklahoma, Texas and Washington. That sounds narrow. It is not small in meaning.
CMS says WISeR is the first Innovation Center model where technology companies are the only model participants. The companies are not selling software around the edge of Medicare. They are embedded in a payment review process, using enhanced technology such as AI and machine learning, with licensed clinicians validating recommendations when payment may be denied.
That is why the new attention from TechCrunch matters. The startup world often looks at health AI through the lens of clinical documentation, patient chatbots or hospital workflow. WISeR points to something more direct: public reimbursement infrastructure as a market for AI-powered utilization management.
The named WISeR participants are Cohere Health in Texas, Genzeon in New Jersey, Humata Health in Oklahoma, Innovaccer in Ohio, Virtix Health in Washington and Zyter in Arizona. These are not consumer apps waiting for patients to discover them. They sit in a process that physicians and suppliers must deal with when certain Medicare services are selected for review.
That gives the model a practical force that venture-backed healthtech companies usually struggle to find. Distribution is one of the hardest problems in healthcare. Selling into providers is slow. Selling into payers is political. Selling into federal programs is even slower, but once a company is inside, the volume and credibility can change the business almost overnight.
CMS frames the model around waste reduction. The selected services include skin and tissue substitutes, electrical nerve stimulator implants and knee arthroscopy for knee osteoarthritis. These are areas where the agency says inappropriate use can create costs for taxpayers and risks for patients, including out-of-pocket bills, complications and anxiety from procedures that may not be supported by the evidence.
According to CMS, WISeR does not change Medicare coverage or payment policy. Providers can seek prior authorization before furnishing a selected service, or they can proceed and face pre-payment medical review. In both cases, the technology companies help assess whether documentation, coverage, payment and coding rules are met.
For founders, that is a serious signal. AI in healthcare is often judged by whether it can pass a clinical safety bar. WISeR adds another test: whether it can enforce rules at scale without turning medicine into paperwork with a better interface.
The incentive problem is just as obvious
The difficult part is how participants are paid. CMS says model participants receive a percentage of expenditures associated with averted wasteful or inappropriate care, adjusted by performance measures that include provider experience. In plain English, the companies can share in savings when services do not get paid.
That structure may reduce unnecessary spending. It may also create suspicion from the first denial. Patients and doctors already have a long memory of prior authorization in Medicare Advantage, where delays and denials have become a political problem. Bringing a related process into Original Medicare, even for a limited list of services, changes the emotional temperature quickly.
KFF recently highlighted that risk in its analysis of WISeR, noting that prior authorization can reduce inappropriate use but can also delay needed care, create uncertainty for patients and add administrative work for providers. KFF found that WISeR services accounted for 5.3% of traditional Medicare Part B spending in 2024, or $12.3 billion, up from 1.1% in 2019.
That growth explains why CMS is interested. It also explains why the model will be watched closely. KFF found that about 207,500 traditional Medicare beneficiaries in the six model states received at least one WISeR service in 2024. The number is not enormous compared with the whole program, but it is large enough for delays and errors to become visible.
Skin substitutes are the most important example. Spending in that category has grown sharply, and CMS has also put a separate nationwide payment change in place for 2026. KFF noted that CMS expects the skin substitute payment changes to reduce Medicare spending on those products by nearly 90% in 2026, which could have a larger effect than WISeR itself in the first year.
That matters because it separates two questions that are often blurred together. One question is whether Medicare has a real low-value care problem. It does. The other is whether AI-assisted prior authorization is the right way to fix it. That answer will depend on speed, transparency and whether clinicians feel the process improves judgment or simply automates friction.
Medicare is becoming the test bed
WISeR is not a chatbot story. It is a payment infrastructure story. The companies involved are being tested on something more consequential than whether an AI tool can summarize a chart. They are being tested on whether technology can sit between a doctor, a patient and a federal payment decision without breaking trust.
If the model works, CMS gets a template for using private technology companies in more parts of Medicare operations. Healthtech startups get proof that AI can do meaningful administrative work in one of the largest healthcare systems in the world. Investors get a clearer view of where the next generation of healthcare automation may actually make money.
If it fails, the lesson will be just as important. A model that saves money while creating denial-delay backlash will not look like innovation to patients waiting for care. It will look like Medicare Advantage problems imported into traditional Medicare with better branding.
That is the line to watch through 2026. The winners in health AI will not be the companies that merely say they reduce burden. They will be the ones that can prove it under real payment pressure, with public scrutiny, clinical accountability and patients who cannot afford to be treated like test data.
Also read: Hollywood backs an AI consent standard that startups cannot ignore • Google is turning the laptop into a Gemini platform • Rivian turns its AI voice assistant into a paid software test
