TLDR: Sumit Rana, President of Epic, is leading the integration of generative AI into more than 100 applications within Epic’s electronic health record (EHR) system. This initiative, highlighted at Epic’s UGM 2025, introduces new AI tools like CoMET, Art, and Emmie, aiming to revolutionize patient care, streamline clinical workflows, and reduce documentation time.
At Epic’s annual User Group Meeting (UGM) 2025, themed around Star Trek, Sumit Rana, President of Epic, unveiled significant advancements in the company’s generative AI strategy. Rana emphasized the critical role of real-world evidence from Epic’s Cosmos database in validating and balancing the outputs of various AI tools, ensuring their safety and reliability in healthcare settings. This year’s presentations showcased a major shift towards a predictive, automated, and collaborative healthcare landscape, with AI simulating future patient health and seamlessly connecting clinicians and patients.
A cornerstone of this initiative is CoMET, a new suite of foundational AI models developed in partnership with Yale School of Medicine and Microsoft Research. CoMET leverages Epic’s extensive Cosmos database, which contains de-identified patient data from over 300 million patients and 16 billion medical encounters. Seth Hain, Epic’s SVP of R&D, stated that CoMET was trained on 151 billion medical events, enabling it to understand the intricate links between events and create comprehensive ‘health care journey storyboards’ for patients. Hain highlighted CoMET’s performance, noting it already matches or surpasses specialized models across 78 different clinical tasks, including predicting diagnoses and disease outcomes. CoMET is slated for release to the Cosmos AI Lab in February 2026, allowing healthcare organizations to test and shape its future applications.
Further enhancing patient and clinician interaction, Epic debuted AI assistants named Art and Emmie. Emmie, designed for patients, will go live in February 2026 with features such as pre-visit agenda setting, simplified lab result explanations, and medication reminders. Art, acting as a physician’s co-pilot in the exam room, utilizes Microsoft’s Dragon technology for ‘ambient scribing.’ This feature, set for an early 2026 launch, instantly pulls data from Cosmos during conversations, automatically captures key details for medical records, and compiles orders and tests for one-click physician confirmation, drastically reducing documentation time.
Also Read:
- Meta Directs Metaverse Division to Achieve Fivefold Productivity Boost Through AI Adoption
- OpenAI’s Sora AI Reshapes Content Creation Landscape, Empowering New Digital Storytellers
Beyond these specific tools, Epic’s CEO Judy Faulkner mentioned that the company is actively working on 160-200 AI projects, a significant increase from the 100 projects reported last year. Rana also outlined four key areas where Epic is leveraging AI to accelerate its internal team’s work: Development & Technical Assistance, Research & Analysis, Communication & Collaboration, and Design & Documentation. He encouraged Epic users to ‘stay curious and learn new ways to use AI,’ ‘reassess risk tolerance for each area,’ ‘know your pioneers,’ and ‘continuously learn and adapt,’ emphasizing the urgency of these adaptations. Rana’s vision extends from ‘Software that assists’ to ‘Software that carries out tasks,’ indicating a long-term goal of transformative automation in healthcare.
