Neil Batlivala has dedicated seven years to creating a healthcare enterprise that remains largely unknown in the tech sector and caters to a patient demographic frequently overlooked by Silicon Valley. However, last month, this endeavor positioned him at the forefront of a much larger initiative.

On April 30, his organization, Pair Team, announced its acceptance into ACCESS, a Medicare initiative — one of 150 selected by the Centers for Medicare & Medicaid Services to explore what AI-enabled healthcare might entail on a national scale. The program is set to launch on July 5.

“The government is establishing frameworks for AI innovation within traditionally regulated sectors,” he mentioned during a Zoom call days later. “The optimal solution prevails, which has not been the case in regulated fields like healthcare.”

ACCESS — Advancing Chronic Care with Effective, Scalable Solutions — is a decade-long CMS initiative testing a payment framework that rewards health outcomes instead of mandatory activities (such as a specific number of check-ins). Organizations like Pair Team will receive consistent payments for managing eligible conditions and will earn the total amount only when patients achieve quantifiable health objectives, like reduced blood pressure or less pain. The program focuses on diabetes, hypertension, chronic kidney disease, obesity, depression, and anxiety.

The payment structure is the significant development.

Traditional Medicare composes reimbursements based on the duration spent with a clinician. There is no system in place to compensate an AI agent that facilitates patient monitoring between appointments, conducts check-in calls, organizes a housing referral, or ensures medication pickup. ACCESS introduces that system for the first time.

“It signifies a transformation in payment models,” Batlivala stated. “Such a thing was impossible before.”

The inaugural cohort includes a diverse array of participants — AI medical startups, virtual nutrition therapy providers, companies with connected devices, and manufacturers of wearables like Whoop. Batlivala expresses skepticism toward some of these entrants.

“I’m an avid supporter of wearables, but for an elderly individual facing food insecurity, I question how much Whoop will actually contribute,” he remarked, adding regarding his own firm, “We’ve been progressing toward this for over five years.”

Pair Team commenced operations in 2019 with a specific patient type in focus: individuals managing chronic conditions while also confronting unstable living situations, food scarcity, or limited access to transportation. Approximately a third of Americans belong to this category.

The foundation of the company was the belief that health outcomes cannot improve without considering the comprehensive context of a person’s life. Currently, it employs around 850 healthcare professionals, operates what it claims to be the largest community health workforce in California, and, according to Batlivala, generates revenue exceeding nine figures. It has secured roughly $30 million in funding from backers such as Kleiner Perkins, Kraft Ventures, and Next Ventures.

The model is supported by peer-reviewed research. A study co-authored by researchers at Pair Team and reviewed by the Journal of General Internal Medicine assessed Pair Team’s community-integrated approach, which merges medical, behavioral, and social care for Medicaid members facing high incidences of homelessness, serious mental health conditions, and chronic illnesses, showing considerable patient engagement and notable decreases in preventable emergency and inpatient usage. Batlivala claims that one out of four hospital visits and one out of two ER visits do not occur when a patient is under his company’s care.

However, for years, providing this standard of care necessitated human teams, which restricted the speed and cost-effectiveness of its expansion. Then, approximately nine months ago, Pair Team implemented a voice AI agent named Flora as its main patient-facing interface. Flora is accessible 24/7, manages intake, coordinates referrals, and ensures follow-ups that keep patients engaged between clinical appointments.

The first call that altered his perspective was with a 67-year-old woman living in her vehicle while coping with PTSD and congestive heart failure. She conversed with Flora for more than an hour. “It was both awe-inspiring and disheartening,” Batlivala recounted. “Flora was likely the only ‘individual’ she had interacted with in weeks concerning her situation.” Now, hour-long dialogues with Flora are commonplace. “That’s the companionship component,” he stated. “And it turns out that it genuinely acts as an intervention.”

The designers of ACCESS are also former startup operators. Abe Sutton, Director of the CMS Innovation Center, and Jacob Shiff, Chief AI and Technology Officer of the CMS Innovation Center, crafted the program. Previously a venture capitalist at a healthcare fund named Rubicon Founders, Sutton has a startup background, as does Shiff, who is a former healthcare entrepreneur. Both joined CMS during the Trump administration, and their startup experiences are evident in the program’s construction: outcome-driven payments, direct-to-consumer enrollment, and a concerted push for competition.

However, there are substantial risks. Participants are sharing extremely sensitive patient information — personal dialogues about housing situations and health issues — within a federal system that has a record of data breaches, including compromised Social Security numbers. For the at-risk demographics ACCESS aims to assist, this is a significant concern.

Financial risks also exist. The success rate of CMS innovation initiatives is mixed. A 2023 analysis by the Congressional Budget Office revealed that the CMS Innovation Center increased federal expenditures by $5.4 billion in its inaugural decade instead of yielding the anticipated savings. CMS is reimbursing less per patient per month than many participants had expected, indicating that the model will only be sustainable for organizations that have largely automated their patient engagement processes.

Batlivala addresses the reimbursement issue by suggesting it is a benefit rather than a drawback. “To construct a model that truly encourages the utilization of AI, the reimbursement rates must be low,” he explained. “The economics are viable only if you’re managing a streamlined, AI-focused operation.”

Pair Team claims it currently holds partnerships that provide access to approximately 500,000 potential patients, with aspirations to reach a million within three years.

Investors in healthcare are observing this closely. Digital health funding reached its peak Q1 total since the onset of the pandemic this year, with AI firms seizing the majority share. Yet, ACCESS has barely made waves beyond the health tech trade media.