ARPA-H تختار Atman Health لتطوير ذكاء اصطناعي صوتي أول للقصور القلبي

The Advanced Research Projects Agency for Health أعلنت on 9 سبتمبر 2026, the teams receiving contract awards under its Agentic AI-Enabled Cardiovascular Care Transformation (ADVOCATE) program, naming Atman Health among the performers selected to build a patient-facing clinical AI agent for heart failure patients. Atman’s سجل الجوائز lists an award of up to $7.7 million.

Atman Health أعلنت عن اختيارها the same day, stating that it will develop an agentic artificial intelligence system for cardiovascular care beginning with heart failure. The award, titled “Safety-First Autonomous Cardiovascular Care: A Hybrid Deterministic-Agentic Architecture,” carries a 9 سبتمبر 2026, start date, names Rahul Deo as principal investigator, lists Needham Heights, Massachusetts, as the location, and is managed by ARPA-H’s Scalable Solutions Office.

The ADVOCATE Program

ARPA-H’s commitment is up to $33.7 million in the first year of the four-year, $62.7 million program. The agency’s stated goal is to create what it describes as the world’s first reliable, FDA-authorized clinical agentic AI system, one that serves around the clock as a new digital member of the clinical care team. If the program succeeds, the agency estimates it will generate $28 billion in annual cost savings across the heart failure population alone.

ARPA-H’s commitment is up to $33.7 million in the first year of the four-year, $62.7 million program. The agency’s stated goal is to create what it describes as the world’s first reliable, FDA-authorized clinical agentic AI system, one that serves around the clock as a new digital member of the clinical care team. If the program succeeds, the agency estimates it will generate $28 billion in annual cost savings across the heart failure population alone.

ARPA-H’s commitment is up to $33.7 million in the first year of the four-year, $62.7 million program. The agency’s stated goal is to create what it describes as the world’s first reliable, FDA-authorized clinical agentic AI system, one that serves around the clock as a new digital member of the clinical care team. If the program succeeds, the agency estimates it will generate $28 billion in annual cost savings across the heart failure population alone.

The agency frames the program around an access problem: more than 200,000 Americans die each year from preventable impacts of cardiovascular disease, the leading cause of death in the country, and nearly half of U.S. counties have no cardiologist at all. On its صفحة البرنامج, ARPA-H says ADVOCATE will support clinical AI agents that can be trusted to autonomously adjust appointments, medications, diet, and exercise, developed in parallel with a supervisory AI “overseer” that monitors deployed agents’ continued safety and efficacy. The program plans deployment with health systems across settings ranging from major health systems to rural clinics.

“Every day, Americans are dying from cardiovascular disease that we know how to prevent and treat — and too often the barrier isn’t the medicine; it’s access to a clinician who can guide the care,” said ADVOCATE Program Manager Haider Warraich, a practicing cardiologist. He described the program’s target technology as a clinician-extender: an autonomous agent able to understand a patient’s treatment needs, provide certain care autonomously, and engage the clinical team when needed.

ARPA-H first أطلقت برنامج ADVOCATE on 13 يناير 2026, as a research and development funding opportunity issued through an Innovative Solutions Opening. The January announcement invited proposals across three technical areas: a patient-facing AI agent, a supervisory component to ensure continued agent accuracy and patient safety, and recruitment of health systems to support design and deployment. It focused the program on patients with heart failure and those who have had heart attacks, defined agentic AI as systems that can act independently to achieve pre-determined goals with minimal human oversight, and described an approach integrating wearable health monitors and digital health records.

Performer Teams Across the Three Technical Areas

In the patient-facing technical area, Tempus AI joins Atman Health with a plan to extend its existing Olivia patient health app, adding continuous monitoring with deeper clinical analysis when the system detects meaningful changes in a patient’s health. Updoc will develop a system whose conversational intelligence is separated from clinical authority by a clinician-built rules system that validates every proposed action against approved protocols before execution.

Stanford University will design the disease-agnostic supervisory agent, which continuously monitors the patient-facing agents after deployment, detecting unsafe recommendations and out-of-distribution behavior in real time. ARPA-H describes a three-stage, compute-escalating pipeline — outlier filtering, rule-based screening, and a deep-research auditing agent — that produces inspectable per-claim rationales.

The implementation technical area covers real-world deployment anchored in health systems. Duke University will provide a multi-site validation platform testing ADVOCATE agents across five health systems and rural sites on both Epic and Cerner/Oracle electronic health records, amplified by the American Heart Association’s national reach. Kaiser Permanente will provide an enterprise-scale deployment across 21 medical centers and more than 260 clinics, embedding the agents into Epic software workflows for heart failure patients and generating reusable deployment blueprints through shadow-mode deployments and pragmatic randomized clinical trials. ARPA-H states this work carries a particular focus on rural and underserved communities where cardiology access is most constrained.

ARPA-H will engage the Food and Drug Administration throughout the program lifecycle to develop the regulatory framework for this new class of patient-facing clinical AI, and selected performers will work with the agency to establish shared evaluation standards, interoperability requirements, and reimbursement pathways. Rick Abramson, director of the FDA’s Digital Health Center of Excellence, said in the agency’s announcement that getting the regulatory framework right requires close, iterative collaboration between developers, clinicians, and the FDA. For independent oversight, ARPA-H has contracted with the Johns Hopkins University Applied Physics Laboratory as its external evaluation partner, assessing technical performance and clinical outcomes.

Atman Health’s Deterministic Engine

Atman says its system separates two functions that most clinical AI systems combine. The large language model interprets records and handles patient conversation by telephone and chat, while clinical decisions are made outside the model by a deterministic engine. The company reports the engine is built on more than 9,000 explicit clinical criteria, 9,600 medication and indication pairings, 6,900 titration profiles, and more than 100 symptom triage flows. It checks exclusion criteria before proposing any medication and records, for each recommendation, the logic that produced it, the evidence behind it, and the hazard classification under ISO 14971. A shadow agent trains and validates the system’s behavior in real-world clinical settings.

“In heart failure, a single wrong adjustment can put a patient in the emergency department with elevated potassium, kidney injury or an arrhythmia,” said Rahul Deo, Atman’s co-founder, chief medical officer and chief product officer. He said that before such a system can make those decisions on its own, a physician must be able to ask what was checked and why, and whether it arrived at the same answer every time.

Co-founder and chief executive officer Rahul Patel said the company chose to encode cardiology as an explicit rules-based system deliberately, so that everything with a clinical consequence happens inside the encoded engine, in a structured form that a cardiologist, client, or regulator can review. ARPA-H’s award description says the Atman system pairs an evidence-based clinical decision engine with large language models behind a voice-first interface that adjusts on the fly to ask each patient the diagnostic questions most likely to be useful. Atman works with provider groups and health plans and is backed by Khosla Ventures and iSelect Fund.

The Treatment Gap and the Regulatory Path

Atman’s announcement describes the clinical problem it is targeting: approximately 7.7 million Americans live with heart failure, a condition treated with four classes of medication that must be introduced and increased gradually over months, with each step checked against kidney function, other lab values, and blood pressure levels. The company describes the regimen as among the most proven in medicine, yet states that in major patient registries only about 1 percent of eligible patients were on target doses of at least three of those foundational drug classes. Citing a 2024 analysis in the Journal of the American College of Cardiology, the company states that 1,454 of the 3,143 counties in the United States have no practicing cardiologist.

The company states that its regulatory path is defined: engagement with the FDA through the pre-submission process in the first year, an Investigational Device Exemption application at month 12, and a De Novo submission targeted within 24 months, preceded by a single-center shadow-mode study. Under the program’s terms, performers in the patient-facing technical area must submit what ARPA-H describes as a first-of-its-kind FDA authorization package for the product within 24 months of contract award.