Clinical artificial intelligence has crossed a critical threshold: approximately 667,000 U.S. physicians—roughly two-thirds of the nation's doctor workforce—now use OpenEvidence's AI-powered platform to query medical literature during patient care. NYU Langone Health's announcement that it has deployed the tool systemwide across its Epic electronic health record system signals that clinical AI has evolved from experimental technology to core health system infrastructure.

The deployment gives more than 6,000 physicians and 54,000 employees across NYU Langone's network—which handles 12 million outpatient visits annually—direct access to AI-generated clinical evidence within their existing workflow. Rather than switching to separate applications, clinicians can now pose natural language questions about patient care and receive answers contextualized against individual patient records.

From Individual Adoption to Enterprise Infrastructure

OpenEvidence's growth trajectory reveals how clinical AI gains traction. A JAMA study tracking usage from January 2024 through June 2026 documented nearly 1 million U.S. clinicians submitting 271 million AI-enabled queries. Monthly active users surged from just 355 in January 2024 to 703,265 by June 2026, while queries per active user climbed from 18 monthly to 39.5—suggesting physicians find increasing value as they integrate the tool into practice.

The 556,200 physicians using the platform averaged 37 queries per month, roughly equivalent to one query per specialist visit in America, which totals approximately 300 million annually. By October 2026, verified U.S. clinicians were conducting 45 million consultations monthly. Month-to-month retention reached approximately 90% in 2026, and by June 2026, 92.5% of active users were returning rather than one-time experimenters.

"The scale and velocity of adoption is remarkable," said Dr. Anupam Jena, corresponding author of the JAMA study and Joseph P. Newhouse Professor at Harvard Medical School. The data shows physicians primarily use the tool for clinical decision-making, with 51.9% of queries related to treatment and management decisions and 13.9% focused on diagnosis.

What Physicians Actually Ask AI

The JAMA study reveals that early-career physicians and specialists in oncology, hematology, internal medicine, and family medicine show the highest usage rates. According to Dr. Jena, physicians tend to consult the AI "where they don't have specific expertise," suggesting the tool addresses knowledge gaps in complex or rapidly evolving clinical domains.

Nader Mherabi, Executive Vice President, Vice Dean, and Chief Digital and Information Officer at NYU Langone, emphasized the integration's workflow advantages. The "patient-aware clinical intelligence" uses what the company calls agentic AI to combine patient record details with medical literature evidence, embedded directly where physicians document care.

NYU Langone joins a growing roster of major New York City academic medical centers deploying OpenEvidence at enterprise scale, including NewYork-Presbyterian (10 hospitals, 350 care sites), Mount Sinai (systemwide), and Memorial Sloan Kettering (precision oncology partnership). This geographic concentration suggests institutional validation and network effects driving adoption.

Unanswered Questions About Patient Impact

Despite the massive adoption documented in the JAMA study, a critical gap remains: patient outcomes. The study notes that whether this AI-assisted clinical decision-making actually improves patient care remains unstudied. The theoretical benefits—more evidence-informed decisions, particularly in complex cases or rare conditions—have yet to be validated through outcomes research.

Daniel Nadler, Founder and CEO of OpenEvidence, has built a platform that demonstrates clear physician engagement and retention. The company's trajectory from bottom-up individual clinician adoption to top-down enterprise contracts represents a viable commercialization pathway for clinical AI tools, with EHR integration proving essential for institutional-scale deployment.

As NYU Langone's 12 million annual visits generate substantial data through this integrated system, the health system is positioned to conduct the outcomes research that could answer whether AI-assisted clinical decisions translate to better patient care—the ultimate measure of success for any clinical technology.

Sources

  1. NYU Langone brings OpenEvidence's clinical AI into the EHR systemwide — Fierce Healthcare