Medicare's Quiet AI Revolution: How the ACCESS Program Is Building the Payment Model That Could Transform Chronic Care — and Why Most of Tech Has Missed It:
Inside Pair Team’s Rise: The $30M Startup Using Voice AI to Solve the Chronic Care Crisis Silicon Valley Ignored.
A 10-year CMS program launching July 5 is quietly rewriting the rules of AI-driven healthcare — rewarding outcomes over activity and making room for technology that traditional Medicare never could.
Introduction: The Healthcare AI Story Silicon Valley Is Sleeping On:
While the tech world has been fixated on AI chatbots, coding assistants, and enterprise productivity tools, a quieter and arguably more consequential AI transformation has been taking shape inside the United States healthcare system — and most of the industry has not noticed. On April 30, a company called Pair Team announced it had been accepted into ACCESS
— Advancing Chronic Care with Effective, Scalable Solutions — a new 10-year Medicare program run by the Centers for Medicare & Medicaid Services (CMS) that is specifically designed to test what AI-driven chronic care management can look like at federal scale. The program goes live on July 5, 2025, with 150 participating organizations. The stakes could not be higher — and the opportunity could not be more underappreciated.
What Is the ACCESS Program?: Medicare's New AI-Ready Payment Model:
The ACCESS program represents a fundamental shift in how Medicare pays for care — and that structural change is the real story behind the headlines. Traditional Medicare reimbursement has always been built around time spent with a clinician. Under the old model, there is simply no mechanism to pay for an AI agent that monitors a patient remotely between visits, makes a check-in call, coordinates a housing referral, or ensures that someone picks up their prescription. ACCESS creates that mechanism for the first time in the program's history.
Rather than rewarding required activities — like a mandated number of check-ins per month — the ACCESS payment model ties reimbursement directly to measurable health outcomes. Participating organizations receive predictable monthly payments for managing qualifying chronic conditions, and they earn the full payment amount only when patients hit concrete health milestones: lower blood pressure, reduced pain scores, better glycemic control, and similar markers. The program covers diabetes, hypertension, chronic kidney disease, obesity, depression, and anxiety — conditions that together affect tens of millions of Americans and account for a disproportionate share of total healthcare spending.
Neil Batlivala, CEO of Pair Team and one of the program's 150 inaugural participants, described the significance of this shift in direct terms. "The government is creating swim lanes for AI innovation in traditionally regulated industries," he said. "The best solution wins — which, in regulated industries like healthcare, that's not been the case." For an industry long defined by rigid billing codes and compliance requirements, that framing is nothing short of revolutionary.
Pair Team: Seven Years Building for the Patients Silicon Valley Ignores:
Pair Team is not a typical AI health startup — and its founder did not build it for the typical health-tech customer. Founded in 2019, the company was built from day one around a specific and underserved patient: people managing chronic conditions who were simultaneously dealing with unstable housing, food insecurity, or lack of transportation. Roughly a third of Americans fall somewhere in this category — a population that traditional health tech has largely bypassed in favor of the more commercially attractive wellness market.
The founding premise of Pair Team was both simple and radical: you cannot meaningfully improve health outcomes without addressing the full social context of a patient's life. That philosophy, now broadly described under the framework of social determinants of health, drove the company to build what it now calls the largest community health workforce in California.
Today, Pair Team employs roughly 850 clinical professionals, has raised approximately $30 million from investors including Kleiner Perkins, Kraft Ventures, and Next Ventures, and generates revenue Batlivala describes as above nine figures annually. The results are backed by peer-reviewed evidence — a meaningful differentiator in a digital health landscape often heavy on claims and light on data.
A study co-authored by Pair Team researchers and published in the Journal of General Internal Medicine evaluated the company's community-integrated care model — which blends medical, behavioral, and social care for Medicaid members with high rates of homelessness, serious mental illness, and chronic disease management. The findings showed strong patient engagement and significant reductions in avoidable emergency and inpatient utilization. Batlivala summarizes the impact starkly: "One in four hospital visits and one in two ER visits don't happen" when a patient is under Pair Team's care.
Flora: The AI Voice Agent Having Hour-Long Conversations With the Forgotten:
For most of its existence, Pair Team's model depended on human clinical teams — which placed real limits on how quickly and cost-effectively the company could scale. That changed approximately nine months ago, when Pair Team deployed Flora, a voice AI agent that now serves as the company's primary patient-facing interface. Flora is available 24 hours a day, seven days a week, handling patient intake, coordinating social service referrals, and conducting the regular check-ins that keep high-risk patients engaged between clinical visits.
The call that changed Batlivala's understanding of what AI could do in this space involved a 67-year-old woman living out of her car, managing both PTSD and congestive heart failure. She spoke with Flora for over an hour. "It was both incredible and depressing," Batlivala recalled.
"Flora was probably the only 'person' she'd talked to in weeks about her situation." Today, hour-long conversations between Flora and patients are routine — not outliers. Batlivala has come to see this dimension of the tool not as a byproduct but as a genuine clinical intervention: "That's the companionship piece. And it turns out that is truly an intervention."
The Hidden AI War
Nobody Is Telling You About
Our latest documentary deep-dive into the geopolitical struggle for machine intelligence dominance. Explore the two paths of AI development: open source vs. closed architecture.
Flora is precisely the kind of AI application that the ACCESS payment model was built to accommodate — and precisely the kind that traditional Medicare reimbursement could never support. An AI agent conducting a 60-minute check-in with a socially isolated patient does not fit into any legacy billing code. Under ACCESS, it can finally be recognized — and paid for — as the clinical work it actually is. This is the connection that makes the CMS ACCESS program so significant for the future of AI in healthcare.
Who Designed ACCESS: Former Startup Founders Now Running CMS Innovation:
The design philosophy behind ACCESS reflects the backgrounds of the people who built it — and those backgrounds are not what you might expect from a federal agency. The program was architected by Abe Sutton, Director of the CMS Innovation Center, and Jacob Shiff, Chief AI and Technology Officer of the CMS Innovation Center. Both joined CMS under the Trump administration. Sutton previously worked as a venture capitalist at Rubicon Founders, a healthcare-focused fund. Shiff is a former healthcare founder.
Their startup DNA is visible throughout ACCESS's architecture — in ways that distinguish it sharply from the typical CMS program. The program features outcome-based payments instead of activity-based billing, a direct-to-consumer enrollment pathway, and a deliberate structural push toward competitive innovation. The first cohort spans a striking range of participants: AI doctor startups, virtual nutrition therapy providers, connected device companies, and wearable technology makers including Whoop. It is, in essence, a federal sandbox for healthcare AI.
Batlivala is candid about his skepticism toward some of the cohort's participants — particularly wearables. "I'm a big fan of wearables, but for a senior who's struggling with food insecurity, I don't know how much Whoop is going to be able to do," he said. It is a pointed observation about the gap between consumer health technology and the realities of the population ACCESS is actually designed to serve — and a reminder that not all AI healthcare innovation addresses the same problem with equal effectiveness.
The Real Risks: Patient Data, Financial Uncertainty, and the CMS Track Record:
No honest assessment of the ACCESS program can ignore its genuine risks — and they are significant. Participants are feeding extraordinarily sensitive patient health data into a federal infrastructure with a documented history of security incidents, including exposed Social Security numbers. The data flowing into ACCESS is not routine — it includes intimate conversations about housing instability, mental illness, and chronic disease. For the vulnerable patient populations the program is designed to serve, healthcare data privacy is not an abstract concern. It is a lived reality with real consequences.
The financial picture adds another layer of complexity. The CMS Innovation Center's track record on cost savings is, at best, mixed. A 2023 Congressional Budget Office analysis found that the Innovation Center actually increased federal spending by $5.4 billion during its first decade — the opposite of its intended impact. Compounding this, CMS is paying less per patient per month under ACCESS than many participants anticipated, which means the reimbursement math only works for organizations that have deeply automated their patient interactions.
Batlivala's response to the reimbursement challenge is counterintuitive — and reveals something important about the program's design logic. "If you want to build a model that truly incentivizes the use of AI, the reimbursement rates have to be low," he argued. "The economics only work if you're running a lean, AI-first operation." In other words, ACCESS is not designed to make the old model work with AI bolted on. It is designed to make the old model financially unviable — forcing a genuine rethink of how chronic care gets delivered.
Scale and Ambition: 500,000 Patients Today, One Million in Three Years:
**Pair Team is entering the ACCESS program **with a significant patient pipeline already in place. The company currently has partnerships that give it access to roughly 500,000 potential patients, and Batlivala has set a target of reaching one million patients within three years. That growth trajectory depends heavily on the ACCESS payment model functioning as designed — and on Flora continuing to perform as a scalable, effective AI patient engagement tool.
The broader investment community is beginning to pay attention to the digital health space again, even if ACCESS itself has barely registered outside health tech trade publications. Digital health funding hit its highest Q1 total since the pandemic this year, with AI healthcare companies capturing the bulk of new capital. If ACCESS delivers on its promise — even partially — it is likely to become one of the most-cited case studies in the intersection of federal health policy, AI adoption, and value-based care.
Conclusion: The Payment Model Is the Product:
The most important thing to understand about the CMS ACCESS program is that the technology is not the breakthrough — the payment model is. AI voice agents, remote monitoring tools, and automated care coordination have existed for years. What has not existed, until now, is a federal payment structure that recognizes and rewards the work they do. That is what ACCESS changes. And that structural shift — not any single piece of technology — is what makes this a potentially generational moment for AI in chronic disease management.
For healthcare founders, investors, and policymakers, the ACCESS program represents exactly the kind of "swim lane" that Batlivala described — a defined, federally sanctioned space where AI can compete on outcomes rather than navigate around billing codes. Whether the program ultimately saves money, improves lives at scale, and justifies the data privacy risks it introduces are questions that will take years to answer.
But one thing is already clear: the federal government has built a payment model for the AI era of healthcare, and the companies that understand that earliest will have a decisive advantage in one of the largest and most consequential markets in the world.
