Recently, I had the chance to sit down with Jacob Shiff, Chief AI and Technology Officer at the CMS Innovation Center, for a conversation hosted by the Health Evolution Roundtable on Integrated Home-Based Care. The topic was ACCESS, the model CMMI describes as Advancing Chronic Care with Effective, Scalable Solutions, and one of the more consequential payment innovations to come out of Washington in years for anyone building technology-enabled care.

Before CMMI, Jacob was the founder and CEO of Anomaly, an enterprise AI company that enables fast and accurate health care payments, so there is a tremendous amount of enthusiasm in the founder / entrepreneurial community about his leadership of AI efforts at CMMI.

I moderate this Roundtable alongside Hal Paz, Mona Siddiqui, and Pippa Shulman, and our group tends to include people who have already lived through a few generations of value-based care experiments. So rather than start with the basics, I asked Jacob to go straight to what makes ACCESS different.

His answer, in short: it is the first CMMI model built to pay organizations for outcomes rather than activities. For years, digital health tools like connected devices, apps, and remote monitoring have been widely available to commercially insured and Medicare Advantage patients, but people in Original Medicare have largely been left out. ACCESS is meant to close that gap. Instead of billing for visits or devices, participating organizations receive monthly payments to manage chronic conditions like hypertension, diabetes, musculoskeletal pain, and depression, and the size of that payment depends on whether patients actually improve.

I asked Jacob what drove CMS to build this now. His view was that the technology has finally caught up to the ambition. AI-enabled monitoring, coaching, and triage tools have matured to the point where they can meaningfully extend a physician's reach into the home, but the payment system never had a lane for that kind of care. CPT codes were built around discrete encounters, not the kind of continuous, largely automated support these tools can provide. ACCESS is CMS's attempt to build that lane.

We spent a good part of the discussion on the operational and financial realities facing organizations that want to participate. This is where the conversation got most candid. Several people in the group pointed out that health systems and independent practices are starting from very different places, and asked how CMS plans to keep the playing field level when infrastructure, data capabilities, and risk tolerance vary so widely. Jacob acknowledged this is a real tension. Without upfront funding, organizations have to be honest with themselves about how much risk they can absorb in the early years, before outcome payments start to materialize. His advice to CEOs weighing participation was to be clear-eyed about their existing infrastructure before committing, rather than assuming the payment model itself will fund the build-out.

One detail that stuck with me is how ACCESS tries to keep primary care physicians in the loop rather than sidelining them. Clinicians can refer Medicare patients with qualifying chronic conditions to technology-enabled care organizations while staying involved in the patient's care and billing for co-management. That structure matters a great deal to primary care providers, where the shared concern is not whether technology can extend care into the home, but whether it does so in a way that strengthens the relationship between patients and their existing physicians rather than fragmenting it.

We also talked about how success will be measured. Traditional value-based care models often rely on utilization and cost metrics that were built for a fee-for-service world. Jacob was direct that ACCESS needs a different scorecard, one built around clinical improvement in the conditions it targets, and that CMMI is still refining what that looks like in practice as more organizations begin participating.

Toward the end, I asked what capabilities organizations should be building now, regardless of whether they join this particular model. Jacob's answer was less about any specific technology and more about organizational discipline: the ability to track outcomes rigorously, the willingness to change course when something isn't working, and the humility to treat early results as data rather than validation.

Since our conversation, CMS has continued to build out the ACCESS infrastructure, and the broader industry conversation about how to pay for AI-enabled care has only intensified. That tracks with what I heard from our Roundtable participants that day. The question is no longer whether technology belongs in chronic disease management. It is how we build payment systems sturdy enough to reward the outcomes that technology can actually produce.

My thanks to Jacob for such a candid conversation, and to the Health Evolution community for pushing the discussion well past the talking points.