Medicare's WISeR Model: AI-Powered Prior Authorization and What It Means for Practices

What a six-year pilot in six states could mean for the future of how Medicare decides what care gets paid for.

On January 1, 2026, CMS launched the Wasteful and Inappropriate Service Reduction (WISeR) Model. It is the first time traditional Medicare has required prior authorization for specific Part B services at scale, with AI-powered review built directly into the process.

The WISeR Model introduces something traditional Medicare has almost never required at scale: prior authorization. It is powered, in part, by artificial intelligence.

If you're enrolled in Medicare Advantage, prior authorization is old news. Private MA plans have used it for years, requiring doctors to get pre-approval before delivering certain treatments. But for the roughly 30 million Americans in Original (fee-for-service) Medicare, this kind of gatekeeping has been largely absent.

WISeR changes that. Depending on how the next six years go, it could reshape the entire authorization landscape for the program.

Key Takeaways

• What it is: The WISeR Model is a six-year pilot (2026–2031) introducing AI-powered prior authorization to traditional Medicare Part B.
• Scope: Currently active in six states (AZ, NJ, OH, OK, TX, WA) targeting 15 specific elective services like knee surgery and nerve stimulators.
• Novel Structure: Private tech firms, rather than insurers, use AI to review requests and are paid a percentage of the savings from averted care.
• The Goal: To curb an estimated $134–$185 billion in annual waste and fraud to help delay Medicare's projected 2033 insolvency.
• The Controversy: Critics warn that the "pay-for-denial" incentive structure and AI-driven gatekeeping could lead to inappropriate care restrictions.

What the WISeR Prior Authorization Model Actually Does

WISeR is a six-year pilot running from 2026 through 2031 in six states: Arizona, New Jersey, Ohio, Oklahoma, Texas, and Washington. The model targets 15 specific Part B services that CMS has identified as having a higher risk of waste, fraud, and abuse. Of these, 13 are currently active and two have implementation delayed to a future performance year.

The active services include arthroscopic debridement for osteoarthritic knee, several categories of nerve stimulators (electrical, vagus, phrenic, sacral, and hypoglossal), epidural steroid injections, cervical fusion, percutaneous vertebral augmentation for compression fractures, incontinence control devices, impotence treatments, and bioengineered skin substitutes for lower extremity wounds. Deep brain stimulation and percutaneous image-guided lumbar decompression were delayed and will be reevaluated for inclusion in a future year. These aren't emergency or inpatient services; they're elective procedures with existing, publicly available coverage criteria.

Here's what's genuinely novel: WISeR is the first CMS Innovation Center model where the primary participants are technology companies, not healthcare providers or insurers. Six private firms have been selected to use AI and machine learning tools to review prior authorization requests. Each company covers a specific state or region and is responsible for issuing determinations, typically within 72 hours.

The financial model is equally noteworthy. These technology participants don't receive a flat fee. Instead, they earn a percentage of the savings generated by averted wasteful care.

In other words, they get paid when services are denied or diverted — a structure that has immediately drawn scrutiny.

Providers aren't technically required to submit prior authorization requests. But if they skip the process and simply submit a claim, it gets flagged for prepayment medical review anyway. Either way, the service is scrutinized before Medicare pays for it.

CMS has also indicated that providers with consistent compliance records may eventually earn "gold-card" exemptions from the process — a carrot designed to ease the administrative burden over time.

Why Now: The Scale of Waste and the Solvency Crisis

To understand why CMS built WISeR, you need to look at what's actually happening with Part B spending. Fee-for-service Medicare pays providers based on the volume of services delivered rather than their value or outcomes, which can incentivize medically unnecessary care. In 2022, the Medicare Payment Advisory Commission estimated that Original Medicare spent between $1.9 and $5.8 billion on low-value services with minimal clinical benefit.¹

By 2024, Part B spending on skin substitutes alone exceeded $10 billion, a nearly forty-fold increase from $256 million in 2019, with the HHS Inspector General flagging alarming trends in utilization, pricing, and outright fraud.² Across the six WISeR states, KFF found that Medicare spent $12.3 billion in 2024 on the specific services now subject to the program. CMS has cited estimates that up to 25% of all U.S. healthcare spending may be wasteful.³

Of that, the two domains most relevant to WISeR — overtreatment or low-value care ($75.7–$101.2 billion annually) and fraud and abuse ($58.5–$83.9 billion) — account for a combined $134–$185 billion in estimated national waste. WISeR targets a narrow slice of that total, but its underlying theory is that AI-assisted prior authorization can address both problems at once: catching low-value care before it's delivered and flagging fraudulent billing patterns earlier in the claims process.

These numbers land at a moment when Medicare's financial future is under serious pressure. The 2025 Medicare Trustees Report projected that the Hospital Insurance (HI) Trust Fund, which finances Part A, covering inpatient hospital care, skilled nursing, hospice, and home health, will become insolvent by 2033, three years earlier than the prior year's estimate.⁴ At that point, incoming revenue would cover only about 89% of projected benefits, forcing automatic cuts to provider payments.

More recent projections suggest the timeline is compressing. The 75-year actuarial shortfall now stands at 0.42% of taxable payroll, up from 0.35% the year before. To close that gap, policymakers would need to either increase the Medicare payroll tax by roughly 14% or cut program spending by about 9%.⁵

Recent tax legislation has compounded the problem: the Committee for a Responsible Federal Budget estimated that the "One Big Beautiful Bill" signed in July 2025 could push HI insolvency from late 2033 to mid-2032.⁶ A February 2026 CBO analysis painted an even grimmer picture under certain assumptions, projecting fund exhaustion by 2040 after factoring in reduced payroll tax revenues and rising per-enrollee costs.

Total Medicare expenditures hit $1.12 trillion in 2024. Against this backdrop, any initiative that promises to curb even a fraction of unnecessary spending carries enormous weight.

The Concerns Are Real

The criticism of WISeR has been swift and bipartisan. Congressional representatives from multiple WISeR states introduced legislation in late 2025 to repeal the model. The American Hospital Association urged CMS to delay the launch by six months.

The House Appropriations Committee approved an amendment to block WISeR funding, though it didn't survive final budget negotiations.

The concerns cluster around several themes.

Incentive structure. Paying technology companies based on denied services creates a structural motivation to restrict care, regardless of medical necessity. CMS counters that performance metrics and provider experience surveys provide accountability.

AI transparency. Several major insurers have already faced Congressional investigations and lawsuits over AI tools that dramatically increased denial rates with minimal human oversight. While CMS requires all WISeR denials to be confirmed by licensed clinicians, the AHA has warned that in practice, human review of AI recommendations is often little more than a check-the-box exercise.

Administrative burden. Studies show 90% of medical group practices already call prior authorization extremely burdensome — and that was before WISeR.¹ For practice administrators managing call volume, referral pipelines, and front desk capacity, a model that flags 25% of covered service claims for prepayment review isn't just a compliance issue. It means more inbound calls from patients asking what happened to their procedure, more staff time documenting medical necessity, and more operational pressure on teams that are already running thin.

Just four days before announcing WISeR, the Trump administration extracted a pledge from major insurers to reduce prior authorization requirements, with HHS Secretary Kennedy declaring that pitting patients against insurers "was not good for anyone." KFF polling underscores the tension: 69% of insured adults say prior authorization is a burden, and 34% call it their single biggest burden — exceeding even cost concerns.⁷

Implications for the Future of Medicare Authorization

WISeR currently covers 13 services in six states, but CMS has built in the ability to expand the model to additional services in future performance years. For practice administrators, that means the authorization workflows you build now — and the tools you use to manage them — need to be designed with that expansion in mind, not just the current scope.

If the pilot demonstrates measurable savings without significant harm to beneficiary access or outcomes, it could provide a template for broader prior authorization requirements across traditional Medicare — effectively importing the managed-care playbook into fee-for-service. CMS would have real-world evidence to justify scaling AI-assisted utilization review nationally.

Conversely, if WISeR leads to widespread inappropriate denials, delayed care, or provider exodus, it could strengthen the growing case for prior authorization reform — the kind ten states pursued through new legislation in 2025 alone.

The stakes extend beyond Medicare. WISeR is a real-world test of whether AI can be deployed responsibly in coverage determinations at government scale. The safeguards CMS has built in — human clinician review of all denials, 72-hour turnaround times, provider experience metrics, and transparency requirements — will either prove that AI-augmented authorization can work fairly, or expose the limits of algorithmic oversight in healthcare.

The Bottom Line

Medicare's WISeR model sits at the intersection of two urgent forces: a program spending over a trillion dollars annually while racing toward insolvency, and a healthcare system already drowning in administrative complexity. The premise — that AI can help identify and prevent wasteful spending before it happens — is compelling. But the track record of AI-driven prior authorization in private insurance is deeply troubling, and an incentive structure that pays technology companies based on denied services raises legitimate questions about whose interests are truly being served.

Over the next six years, WISeR will generate data that could shape Medicare policy for a generation. The outcome will come down to execution and oversight. The key question: can CMS drive real savings without creating the denial and delay problems that have eroded trust in Medicare Advantage prior authorization?

For practices already stretched thin on phones and front-desk bandwidth, WISeR adds a new layer of prior authorization burden. Expect more patient calls asking why a procedure was denied. Staff will spend more time navigating authorization workflows and managing documentation pressure before care can be delivered.

This is exactly the kind of compounding administrative load that AI phone agents and automated intake tools are built to absorb. They free your clinical team to focus on care, not calls.

For practice administrators and clinical leaders managing these workflows, the next six years will require close attention. How WISeR performs will shape what prior authorization looks like across traditional Medicare for a long time.