A Federal Payment Model Built for Software
On July 5, one of the more consequential experiments in American healthcare technology quietly went live, and most of the enterprise technology world has yet to register it. The Centers for Medicare and Medicaid Services launched the ACCESS Model, short for Advancing Chronic Care with Effective, Scalable Solutions, a ten year test that pays for chronic disease management based on whether patients actually get healthier rather than on the volume of services billed. More than 150 participants were selected to run it, and the roster reads like a who's who of digital health.
We think ACCESS deserves the attention that hyperscaler capex and frontier model launches routinely command, because it changes the economics that have long strangled health technology adoption. For years the barrier to deploying remote monitoring, AI diagnostics and virtual care was not capability but reimbursement. If Medicare will not pay predictably for a software driven intervention, hospital systems and startups cannot build durable businesses around it. ACCESS is Washington attempting to fix that at the level of the payment code.
How Outcome Aligned Payments Work
The mechanism is a new construct CMS calls Outcome Aligned Payments. Participating organizations receive stable, recurring payments for managing qualifying conditions such as diabetes, hypertension, chronic kidney disease, obesity, depression and anxiety, but they earn the full amount only when patients hit measurable goals like lower blood pressure or reduced pain. The design deliberately inverts fee for service, which rewards activity, and replaces it with a model that rewards results, shifting financial risk onto the technology providers who claim their tools work.
That risk transfer is the whole point, and it is why the participant list matters. Verily, the Alphabet health unit, sits alongside Noom, WeightWatchers, brain health platform Isaac Health, telehealth provider HealthTap, AI doctor startup Doctronic, remote monitoring firm Withings, and mental health providers Headspace and SonderMind, among others. These are companies that have spent years arguing their software improves outcomes. ACCESS hands them a federal stage to prove it, and a payment stream if they can, but exposes them if the outcomes do not materialize.
Where the AI Actually Lives
CMS is explicit that ACCESS is meant to nurture a technology enabled ecosystem, and artificial intelligence threads through it in ways that are practical rather than speculative. The agency points to AI diagnostics that identify people with conditions likely to benefit from the program, connected devices that monitor biomarkers between visits, and software that streamlines the administrative workflows clinicians drown in. The common thread is that the AI has to earn its keep against a health outcome, not a demo metric.
This is a materially different bar than the one most enterprise AI pilots face. Inside corporations, an agent that drafts emails or summarizes tickets can coast on soft productivity claims. Under ACCESS, an AI that flags a hypertension patient is worth money only if that patient's blood pressure comes down and the payment triggers. We suspect that constraint will separate the vendors with genuine clinical validity from those selling dashboards, and it will do so with real dollars rather than analyst commentary.
The Modernization Lesson for Every CIO
There is a broader digital transformation lesson here that reaches well beyond healthcare. ACCESS is, at its core, a governance and incentives redesign that makes technology adoption rational by tying payment to outcomes. Enterprise CIOs wrestling with stalled AI pilots face a structurally similar problem: initiatives fail not because the models are weak but because no one owns the outcome and no budget line rewards achieving it. The federal government is modeling a fix, and private sector leaders should study the structure.
The uncomfortable corollary is accountability. When payment depends on measurable results, the organization has to instrument everything, define what success means before deployment, and accept that some initiatives will not pay off. That discipline is exactly what has been missing from the wave of AI experimentation that produced impressive proofs of concept and disappointing production returns. ACCESS forces the question that too many transformation programs dodge: if this technology works, show the number that proves it, and let the payment follow.
The Headwinds Are Real
None of this guarantees success, and the skeptics have a case. Early reporting flagged that the ACCESS payment rates are modest and hedged with restrictions, though the constraints did not discourage digital health companies from applying in droves. A ten year model is also a long bet in a policy environment that can shift with administrations, and outcome measurement in chronic care is genuinely difficult, with confounding factors that make it hard to credit any single intervention for a patient's improvement.
There is also the question of whether outcome based payment inadvertently pushes participants toward healthier, easier to manage patients, the risk selection problem that has haunted value based care for a decade. CMS will need robust guardrails to prevent the model from rewarding cherry picking rather than genuine population health gains. We flag these not to dismiss ACCESS but to set expectations. This is an experiment, and the honest measure of it will not arrive for years, in the form of outcomes data rather than launch day optimism.
Why It Matters Beyond Healthcare
For technology executives watching from other industries, ACCESS is a signal about where large scale AI adoption is heading. The most credible path to deploying autonomous and assistive systems at scale runs through accountability structures that pay for results, not activity. Healthcare, with its life or death stakes and its enormous public payer, is becoming an unlikely proving ground for that principle, and the lessons will travel to finance, insurance and government services that share the same outcome measurement challenge.
We will be watching the participant cohort closely, because their results over the coming years will either validate outcome aligned payment as a durable engine for technology adoption or expose it as another well intentioned pilot that could not survive contact with clinical reality. Either way, the model deserves a place on the enterprise radar. When the federal government restructures how it pays for software driven care, it is not just reforming Medicare. It is writing an early draft of how institutions will buy AI when they finally insist on being shown that it works.
