VERONA, Wisc.—Epic is ramping up its efforts to infuse artificial intelligence throughout its electronic health record platform, unveiling Tuesday a slew of new AI features and tools.
The EHR giant is integrating generative AI, AI agents and other capabilities broadly across its platform, including clinical documentation, revenue cycle management, patient-facing solutions, operational workflows and clinical trial management.
This year's annual meeting featured a sci-fi theme, with plenty of "Star Trek" references sprinkled in.
The company is working on more than 160 AI projects, CEO Judy Faulkner told the audience of thousands of Epic customers at the company's annual Users Group Meeting. "Some of them are completed or in process, and some of it will be permeating every area of our software," she said.
"We are combining the intelligence and curiosity of the human being with the investigative capabilities of gen AI. And then for the data, there's the EHR, the patient's medical record and integrated with the patient's MyChart. Then there is the Health Grid, plus the learnings from the huge amount of high-quality patient data in Cosmos. Instead of calling it artificial intelligence, we're calling it healthcare intelligence. We think these are going to make significant changes," Faulkner said.
Epic's Health Grid feature enables providers to connect with the broader healthcare ecosystem including payers, specialty diagnostic labs, medical devices and telehealth companies. The company's Cosmos research database encompasses 300 million patient records from more than 16 billion encounters across four countries, Faulkner said Tuesday.
The company offers a growing list of AI features, with many more to come. AI features coming next year include a discharge planning agent, patient-reported outcomes insights, a presurgical risk calculator, a MyChart virtual assistant, patient-facing imaging results overview and a patient flow agent, just to name a few.
Epic is focused on building out gen AI tools to help clinicians and patients while also helping healthcare organizations "stay financially strong," Faulkner said.
The company officially announced its previously rumored AI scribe solution, called Art for Clinicians, along with a gen AI tool for revenue cycle, called Penny, to help with coding and denial appeals. It also unveiled an AI solution for patients, called Emmie, to help patients schedule visits and then prepare for their medical visits.
Emmie can explain results with context, answer open-ended questions to guide patients through healthcare and take actions like scheduling visits or recommending health screenings, according to the company. The clinician-facing AI tool, Art, then anticipates what’s needed for the visit and tees up tailored patient summaries for physicians. Actions taken during the patient visit like orders and documentation are added to a “shopping cart” to be reviewed and signed at the end of the visit, executives said.
Art can also tap into the Cosmos platform for diagnostic insights and to identify patients with similar conditions and diseases.
Epic is collaborating with longtime partner Microsoft to develop its AI charting technology, Faulkner said. Doctors will use Epic's Haiku and Canto to record patient conversations, while the company will use Microsoft's Dragon Ambient AI technology for transcription, she said. Epic then takes that information to create the final note and visit documentation.
"Epic’s native AI Charting is being developed, and we anticipate that it will be ready for limited use by early next year," an Epic spokesperson said.
Healthcare organizations can continue to use third-party AI scribe vendors as well, according to Epic.
"At Microsoft, we believe AI can shape a future for all people on the planet to live healthier. Our ongoing collaboration with Epic supports our vision to help transform the healthcare experience through innovation that delivers meaningful outcomes," Joe Petro, corporate vice president of Microsoft health and life sciences solutions and platforms, said in a statement.
In March, Microsoft unveiled its Dragon Copilot, a clinical AI assistant designed to streamline workflows and enhance the clinician-patient experience.
"We’re proud to be collaborating with Epic to explore how we can bring our core Dragon ambient AI technology to Epic’s new AI Charting capability to further improve care delivery," Petro said.
On the patient-facing side, the company announced the upcoming MyChart Central feature that will let patients use a single Epic-issued ID to connect their MyChart records across different providers. Epic anticipates broad rollout of MyChart Central in November.
"This is a huge time of growth for technology and medicine, things that couldn't be done before," Faulkner said.
RCM and operational tools
Epic has developed gen AI tools to tackle revenue cycle management and reduce administrative burden with AI-based coding and AI-generated denial appeal letters, and these features are already available, Faulkner said.
Advancements with AI make it possible to parse through unstructured data, including provider notes, to improve coding and billing, noted Ryan Krause, a product informatics executive at Epic, during a session with reporters after the executive address.
"There's a lot of payer hoops that our customers have to jump through to make sure that their claims get paid. We're at this really interesting point now where we've got generative AI that can read both the structured and unstructured data to make our customer sites much more efficient in things like coding and getting those bills out the door and getting them paid," Krause said.
The company also is stepping up its work with payers. This summer, Epic hosted a small group of CEOs from both health plans and health systems, according to Seth Howard, vice president of R&D at Epic.
"We walked away with a clear set of shared priorities to improve prior authorization, automate claims management and make medical policy more transparent," he said during the executive address. "Part of the challenge here is the fragmented technology on the health plan side."
To solve these challenges, Epic offers its Tapestry software suite, which enables insurers to track enrollment, authorize services, manage high-risk members, pay claims and capitation and engage members all within a single system. A growing list of payers have signed on to Tapestry, including, just this year, Humana, which marks the first national health plan to do so, Howard said.
As Epic continues to expand its capabilities beyond the EHR, it's developing an enterprise resource planning (ERP) suite, a product that executives said was in the works during HIMSS 2025 back in March.
The ERP will work "hand in glove" with Epic's EHR, Faulkner said, and will feature modules for workforce management including staffing, scheduling and credentialing, supply chain for inventory management and procurement, and financial operations.
Additional services in the works
Epic plans to launch an integrated clinical trials management system that will be available to early adopters next year.
"You've asked us to do this for years, but we've said there are other systems available and you don't need us, but now we've gotten lots of feedback that you want more than the other systems offer," Faulkner said.
The company also is working on an outbreak monitoring and detection system. "Using Cosmos, we're going to create a system to detect outbreaks and notify healthcare organizations," Faulkner said.
"As technology rapidly evolves, we feel a tremendous sense of urgency to transform what might feel like science fiction into real, practical solutions," Howard told the audience.
Cosmos combined with AI for health predictions
Epic is tapping into its massive Cosmos research database to develop predictive tools, such as its Cosmos Median Length-of-Stay for Discharge Planning tool.
The company also previewed Cosmos AI, a family of generative models trained on 115 billion medical events to simulate patient trajectories and predict patients' future health risks.
"By combining the medical language, we created the latest computer science techniques and the breadth and depth of Cosmos, a new family of foundational models emerged," Seth Hain, senior vice president of R&D at Epic, said Tuesday.
The foundation models were evaluated on 78 real-world tasks spanning diagnosis prediction, disease prognosis and healthcare operations.
"We started by comparing to traditional predictive models like a 30-day readmission model, and we built multiple Cosmos AI models, each learning from more and more patient data. As theorized, they improved, ultimately performing as well or better than those traditional models, and that is just one of 78 tests we've run so far," Hain told meeting attendees.
"To be clear, one model built from scratch can predict all of this, and that's the most fascinating thing to me. We never explicitly taught Cosmos AI what a readmission or a lab was. We simply trained it to predict the next event in a patient story over and over and over and it learned that these underlying relationships between intervention, treatments and outcomes, and across all of those tests. We're finding that it continues to improve as it sees more patients, and having only used a billion encounters so far, we're just getting started," Hain said.
Epic collaborated with Yale University and Microsoft Research on a paper detailing Cosmos AI's foundation models. Cosmos AI will be available to researchers and data scientists to run tests on the AI models, Epic executives said.
Editor's note: This article has been edited to accurately reflect that the Cosmos platform encompasses 16 billion patient encounters.