On April 14, CMS announced that more than 150 organizations had been accepted to participate in the ACCESS Model. A 10-year payment model. Launching July 5, 2026. Outcome-aligned payment framework. Private payers representing 165 million members have committed to aligning with the model's payment approach. Eleven weeks from now, it goes live.
It was not the loudest announcement of the week. Most health system leaders I spoke to this week had not seen it yet, or had seen the headline but not the substance.
The substance is this: CMS has published the reference framework for chronic care payment for the next decade. If your health system has clinical AI tools in chronic care management, those tools now have to justify themselves against a specific outcome framework, on a specific timeline, with a specific set of conditions in scope.
CMS also extended the initial application deadline to May 15 to accommodate more participants at launch. Advancing Chronic Care with Effective, Scalable Solutions. That is what ACCESS stands for. The design intent is in the name.
The model focuses on the conditions that drive more than two-thirds of Medicare beneficiary spend: high blood pressure, diabetes, chronic musculoskeletal pain, and depression.
Participants take on downside risk for outcomes at the patient-population level. The payment design is outcome-aligned, not volume-aligned. It pays for population-level movement on the conditions CMS has chosen to measure, not for visits, documentation, or encounters.
This is a meaningful departure from how most health systems have been thinking about the ROI of their clinical AI spend.
In the pilot-and-deployment mindset, an AI tool earns its keep if it reduces documentation time, improves clinician satisfaction, and generates clean utilization data. Those are internal-facing metrics. They matter to operations. They do not show up in ACCESS.
In the outcome-aligned mindset, an AI tool earns its keep if it produces a measurable directional shift in a population-level metric CMS is going to pay for. That is a different ROI calculation. The denominator changes. The evidence requirements change. The vendor conversation changes.
"Your AI tools do not have a scale problem. They have an outcomes problem."
Most health systems I work with have 8 to 15 chronic care AI tools running in parallel. Each was approved by a different committee, each has its own vendor pitch, each has its own internal champion.
Ninety percent of those tools have never been scored against the population-level outcome framework their organization is about to be paid against. The portfolio review is not a nice-to-have. It is the starting position for ACCESS participation.
An annual AI ROI review is the standard at most health systems. Under ACCESS, that is not frequent enough. Outcome-aligned payment needs a quarterly governance cadence, minimum, that tracks tool-level contribution to the population metrics that are driving the payment.
The reporting structure most health systems have in place today was designed for internal operations reporting. It was not designed for payment-facing performance tracking. The structural gap is not about data availability. It is about what the governance body is organized to answer.
Every vendor pitch on a chronic care AI tool has been built around the pilot-deployment ROI frame. Documentation reduction. Clinician satisfaction. Utilization improvement. The vendor conversation that needs to happen now is different: can you show me, at the patient-population level, that your tool moves the outcomes ACCESS will pay for?
If not, how would you design the measurement? Who would validate it? Most vendors do not have a good answer to this question yet. That is useful information.
If your CFO is reading the same industry news you are, you are about to get asked two questions.
First: which of our chronic care AI tools contribute to ACCESS-measured outcomes?
Second: which ones do not?
The answer to the second question is going to get short very quickly under outcome-aligned payment. Tools that are contributing to internal operations metrics but not to payment-facing outcome metrics become budget-line candidates when the organization is on the hook for downside risk.
This is not a case for killing all of those tools. Some of them may have real operational value that justifies their keep even if they do not map directly to ACCESS outcomes. The case is for being able to defend each one on the specific terms the organization is about to be paid against.
Most health systems cannot defend the portfolio on those terms today.
The strategic question is not "which AI tools should we buy next."
The strategic question is "does our governance structure aim AI at the outcomes that pay?"
Most health systems have a procurement process, a clinical evaluation process, and an implementation process. Each of those processes answers a narrower question: vendor risk, clinical validity, operational fit. What most health systems do not have is a strategic governance structure that maps every AI tool in the portfolio against the outcome framework their organization is going to be paid against.
That gap is not a technology gap. It is a structural gap.
The health systems that will win under ACCESS are not the ones with the best AI tools. They are the ones whose governance structure can point their AI tools at the right outcomes.
This is what an implementation advisory retainer actually buys you in Q2 of 2026. Not vendor selection. Not pilot design. Portfolio governance against the payment framework that is about to go live.
If the ACCESS Model announcement is on your desk and you are trying to figure out what it means for your chronic care AI roadmap, that is the conversation I have every week with health system executives.
Ten weeks from now, you will wish you had started the portfolio audit in April.
Portfolio audit, governance cadence design, vendor conversation scripts, executive briefing. Built for health systems preparing for July 5, 2026 launch.
Review the advisory scope at drsarahmatt.com/consultingIf you want to talk through your specific portfolio, book a discovery call.
calendly.com/sarahmattmdTHE SARAH MATT BRIEFING
The practitioner-level context doesn't live on this blog.
Each week, subscribers get the field specifics, vendor questions, and practitioner tools that go beyond the public post. This week: the five questions to put to your chronic care AI vendors right now, and the internal portfolio scoring sheet we are using with health system clients this quarter.
Weekly. No filler. Unsubscribe anytime.
Subscribe at drsarahmatt.comAnd if you want the long-form case on pilot-to-implementation failure modes, Episode 15 of The Clinical Realist drops Wednesday: "The Pilot Trap." The ACCESS Model will reward the health systems that closed the pilot-to-implementation gap. Episode 15 covers why most have not.