The WISeR Model: What Medicare Agents Need to Know in 2026

Your client has had Original Medicare and a Medigap plan for six years, and you’ve told her, truthfully, that Original Medicare doesn’t make her jump through prior authorization hoops the way a Medicare Advantage plan might. Then she calls, confused, because her doctor’s office in Columbus just told her a nerve stimulator procedure needs something called a “WISeR” review first, and she’s never heard the term and neither, maybe, have you. The WISeR model is real, it’s active right now in six states, and it’s the first time CMS has bolted AI-assisted prior authorization onto traditional fee-for-service Medicare at any real scale. This article covers what it actually requires, which states and services it touches today, what it costs when nobody explains it, the manual steps to find out if a specific client is affected, and where Ambrose’s War Room can take the research work off your plate.

What the WISeR model actually is

CMS’s own framing is unusually direct: waste in health care, the agency says in its provider factsheet, “can not only harm patients but also contributes to up to 25% of health care spending in the United States” (CMS, WISeR Model Provider Factsheet). The WISeR model is CMS’s attempt to do something about a narrow slice of that, in Original Medicare specifically, by pairing six named technology companies with the Medicare Administrative Contractors (MACs) that already process claims in each region.

The structure is simple once you see it laid out. CMS picked a short list of services it considers “vulnerable to fraud, waste, and abuse,” picked six states, and assigned one technology company to each state’s MAC jurisdiction:

The timeline is set: CMS published its request for applications June 27, 2025, applications were due July 25, 2025, the model launched January 1, 2026, and it runs through December 31, 2031 (CMS, WISeR Model Provider Factsheet; CRS Report IF13133). It’s being run under the CMS Innovation Center’s authority to test “innovative payment and service delivery models to reduce program expenditures” under Section 3021 of the Affordable Care Act, which is also the authority CMS is using here to waive the statutory limits that have historically kept prior authorization out of traditional Medicare (CRS Report IF13133).

Here’s the mechanical detail that matters most for how you talk to a client: a provider or supplier can choose to submit a prior authorization request before performing the service, or skip that and go through post-service, pre-payment review instead. Either way, “there will be no change in provider payments or providers’ appeal rights” (CMS, WISeR Model Provider Factsheet). The technology companies aren’t adjudicating in a vacuum, either — clinicians employed by the participating companies review every non-affirmation, and a human clinician reviews every denial (CMS, WISeR Model Provider Factsheet; CRS Report IF13133).

Why this breaks a pitch you’ve probably made for years

If you’ve sold Medicare for any length of time, you’ve said some version of this sentence: “With Original Medicare, you go to any doctor who takes Medicare, no referrals, no prior authorization, no network.” That sentence has been true, in practice, for almost the entire history of the program. The Congressional Research Service calculated that in FY2023, traditional Medicare’s existing prior authorization programs — the narrow set that covers things like certain cosmetic procedures, non-emergent ambulance transport, and durable medical equipment — reviewed 3.1 million claims, representing less than 1% of the 1.2 billion total Part A and B claims filed that year (CRS Report IF13133). Prior authorization in Original Medicare has been a rounding error.

Medicare Advantage is the opposite story, and it’s worth knowing the actual scale of that contrast, because it’s the thing CMS is explicitly trying to borrow when it says WISeR “can apply artificial intelligence technologies that MA plans are already using to streamline PAs” (CRS Report IF13133). KFF’s most recent annual analysis found Medicare Advantage insurers made nearly 53 million prior authorization determinations in 2024 — about 1.7 per enrollee — of which 4.1 million, or 7.7%, were fully or partially denied. Only 11.5% of those denials were ever appealed, but when they were, 80.7% were partially or fully overturned (KFF, Medicare Advantage Insurers Made Nearly 53 Million Prior Authorization Determinations in 2024, published Jan. 28, 2026).

That gap is exactly why “Original Medicare doesn’t do prior authorization” has been such a durable, true-sounding line for so long, and exactly why WISeR is a meaningful crack in it rather than a cosmetic change. It’s still a narrow pilot — fourteen service categories, six states, and CMS states explicitly that “codes in WISeR do not overlap with existing Medicare prior authorization programs” (CMS, WISeR Model Provider Factsheet) — but it’s the first time the fee-for-service side of the program has used AI-assisted review at any scale worth a client noticing.

The services covered right now

CMS’s WISeR Model Provider Factsheet lists the service categories in scope for Performance Year 1, each tied to a specific National or Local Coverage Determination:

That last row has its own asterisk worth knowing: the skin and tissue substitute category only applies “in selected WISeR MAC jurisdictions and states with an active LCD in place during the WISeR PYs starting on January 1, 2026, as of November 21, 2025” (CMS, WISeR Model Provider Factsheet). CMS also says plainly that these codes and the clinical documentation requirements behind them “may be subject to change,” and the authoritative, current list lives in the WISeR Provider and Supplier Operational Guide, not in any article, including this one (CMS, WISeR Model Provider Factsheet).

What it costs when nobody explains this to a client

There’s no WISeR-specific denial data yet — the model only launched in January 2026, and CMS hasn’t published outcome statistics for it. What you can source, and what’s directly relevant to how a client experiences this, is what already happens industry-wide once AI-assisted prior authorization enters a process. KFF’s 2024 Medicare Advantage figures are the closest documented precedent for how an AI-assisted review pipeline tends to behave: a 7.7% denial rate, and — this is the number that should worry you more than the denial rate itself — only 11.5% of denials ever get appealed, even though 80.7% of the appeals that do happen succeed (KFF, Medicare Advantage Insurers Made Nearly 53 Million Prior Authorization Determinations in 2024).

Read those two numbers together and the real cost comes into focus. It isn’t that AI-assisted prior authorization produces a flood of permanent denials. It’s that most people who get denied never appeal at all, even when the odds of winning are better than four in five. That gap between “denied” and “appealed” is where an agent’s explanation either closes or widens, depending on whether the client understands that a denial is a starting point, not a final answer.

What it costs you directly, as the agent, isn’t a dollar figure CMS publishes. It’s the phone call that starts with “you told me this wouldn’t happen,” and the ten or fifteen minutes you now spend explaining a federal pilot program you didn’t create and can’t control. That conversation goes a lot better when you’ve already looked up whether the client’s specific service and state puts them in scope, instead of learning about WISeR from the client.

The manual method: finding out if a client is actually affected

None of this requires anything beyond a web browser and the client’s upcoming procedure. Here’s the complete process.

What agents should never do with this

Keep your role narrow, on purpose. You are not the provider, you are not a clinician, and you are not the client’s lawyer. If a client calls about a WISeR denial, your job is explaining what the program is, confirming the appeal path exists, and pointing them to the provider’s billing office or 1-800-MEDICARE for the actual appeal filing — not drafting the appeal letter, not interpreting the clinical rationale behind a non-affirmation, and not promising an outcome. “Results may vary,” and that’s not boilerplate here: CMS’s own data shows most appealed MA denials get overturned, but “most” isn’t “all,” and you have no way to know which bucket a specific case falls into.

Watch the compliance line, too. Walking a client through what WISeR is and how appeals work is servicing an existing policyholder, not Medicare marketing. The moment that conversation drifts into “have you thought about switching to a Medicare Advantage plan instead, since they handle this differently,” you’ve crossed into a Third Party Marketing Organization (TPMO) activity, with the Scope of Appointment and TPMO disclaimer requirements that come with it under CMS’s Medicare Communications and Marketing Guidelines. Keep the two conversations separate, even in the same phone call.

If AI tools of any kind touch this workflow, the NAIC’s AI Model Bulletin, adopted in December 2023, is the governing reference: it “reminds insurers that decisions or actions made or supported by AI must comply with all applicable insurance laws and regulations, sets forth expectations as to how insurers will govern the use of AI, and advises insurers of the type of information the Department may request during an investigation or examination” (NAIC, Insurance Topics: Artificial Intelligence). States adopt it individually and adoption is ongoing, so confirm your own state’s current status before assuming it applies the same way everywhere.

Where this goes next: a live fight, not a settled question

The honest answer to “is this going to spread to my state” is that nobody knows yet, and both directions have real momentum behind them right now. On one side, Abe Sutton, the director of the CMS Innovation Center, told a Paragon Health Institute webinar on September 3, 2026 that WISeR “is actually something that should serve as a model for other lines of business in how to approach authorization for services in a way that is rapid — within 72 hours — in a way that is based off of clinical evidence” (HFMA, CMMI head: WISeR should serve as model). That’s the head of the agency running the pilot saying, in public, that he wants to expand the approach.

On the other side, the House Appropriations Committee’s FY2027 Labor-HHS-Education bill, advanced at a markup on June 9, 2026, includes language that would bar CMS from using any funds “to implement the Wasteful and Inappropriate Services Reduction (WISeR) model or another model that seeks to use prior authorization practices in the traditional Medicare program” (U.S. House Committee on Appropriations, Full Committee Markup: FY2027 Labor-HHS-Education; Becker’s Payer Issues, House committee votes to block funding for WISeR prior authorization pilot). That provision still has to survive the full House, the Senate, and a presidential signature before it means anything, and appropriations riders like this one frequently don’t survive conference. The Congressional Research Service separately notes that members of Congress have sent CMS letters raising concerns, legislation has been introduced to prohibit the model outright, and litigation from providers or beneficiaries is a live possibility (CRS Report IF13133).

Nothing about this changes what’s true today: WISeR is active, in six states, for a specific list of services, right now. What it means for your clients is that this is worth a calendar reminder to check back on, not a one-time read. A rule that’s this contested in Congress and this openly championed by the agency running it is a rule that’s likely to look different a year from now, in one direction or the other.

Where the manual version breaks down at scale

Everything above — the state check, the service match, the operational guide lookup, the appeal-path explanation — is a five-minute task for one client. It becomes a different problem when you’re trying to stay current on this alongside every other CMS rule change that lands between now and AEP: the SOA timing changes, the marketing claims documentation changes, the Part D redesign, whatever shows up in the next Federal Register notice. Nobody has the bandwidth to manually re-check appropriations.house.gov, congress.gov, and CMS’s Innovation Center page every week for every program that might touch their book. That’s not a discipline problem. It’s a research-capacity problem, and it’s the same one this article’s own sourcing ran into: a CMS factsheet, a CRS report, a committee markup page, and a trade outlet, cross-checked by hand, to answer one question.

How Ambrose’s War Room and the Brain actually help with this

This is the part worth being precise about rather than oversold. Ambrose doesn’t have a dedicated “WISeR tracker” feature, and this article isn’t going to claim one exists. What’s documented and real: Ambrose’s architecture describes the Brain as “an internal service that fronts 25+ federal/healthcare data MCPs (CMS, NADAC, FDA, Federal Register, FEMA, …)” (Ambrose docs, Architecture). Sam Okafor, the War Room’s Research head, is documented with access to “brain (all .gov MCPs), marketplace-finder, agent-vault,” and the architecture notes that heads can dispatch their own depth-2 specialists via the Task tool, giving the exact example “Sam → Federal-Register specialist for rule lookups” (Ambrose docs, Executive Team).

Put plainly: asking the War Room something like “what’s the current status of the WISeR model, and has Congress actually blocked its funding yet” is precisely the kind of question that dispatch path is built to answer, by pulling the Federal Register and the Brain’s other federal sources instead of a person doing it by hand. That’s the comparison worth sitting with: the research in this article — a CMS PDF, a CRS report, a committee’s own markup page, a trade outlet’s reporting on a named official’s quote — is exactly the kind of multi-source federal lookup the War Room’s Research head is designed to run on request, in the time it takes to ask a question instead of the time it took to write this section.

What you get by joining

One Ambrose seat, including the War Room and its Research head, comes with a Tech Savvy Insurance membership: $97 a month, billed monthly, cancel anytime, founding rate locked in while the membership stays continuously active. Alongside the seat: weekly Zoom calls with open Q&A and build-with-you sessions, more than 30 hours of recorded training updated monthly, Meta Ads and AI marketing training built for health and life agents, pre-built AI templates and bot deployments, and a free annual in-person member workshop — in an explicit no-recruiting zone, so you can ask a real compliance question without getting DM’d about a downline an hour later.

Everything in this article — the state check, the service list, the operational guide, the appeal path — works whether you ever join anything or not. Go run it by hand on the next client who mentions one of these procedures, or let the War Room run the lookup for you.

Related reading: our guide on Medicare drug cost lookups for the Part D side of coverage denials, our breakdown of AI compliance for insurance agents for the NAIC Model Bulletin in full, and our look at Medicare Advantage ghost networks for the network-side version of this same “what the client was told vs. what’s actually true” problem.

The close

CMS didn’t build WISeR to catch you off guard. It built it to test whether AI-assisted review can trim a slice of what it calls wasteful spending in a program that’s otherwise run prior-authorization-free for sixty years, in six states, for fourteen service categories, with appeal rights fully intact. The problem for agents is that nobody sent a memo to the field, and “Original Medicare doesn’t do prior authorization” is a sentence a lot of us have said for years without an asterisk. Add the asterisk, bookmark the state and service list, and you’ve covered the entire client-facing risk. This is the kind of live regulatory tracking we work through on a Tuesday call, watching the War Room’s Research head pull the current status instead of six browser tabs. $97 a month, cancel anytime, and nobody will pitch you a downline: https://techsavvyinsurance.com/.