226: Epic's CMO Dr. Jackie Gerhart: Epic's Vision for The Intelligent Visit, Which of Epic's 100+ AI Use Cases Get the Most Adoption, and What Changes About Physician Identity in the Age of AI

Subscribe on: RSS | SPOTIFY | APPLE PODCAST | GOOGLE | BREAKER | ANCHOR

On this episode of The Digital Patient, Dr. Joshua Liu, Co-founder & CEO of SeamlessMD, and colleague, Alan Sardana, chat with Jackie Gerhart, MD, Chief Medical Officer at Epic Systems, about "Epic's Vision for The Intelligent Visit, Which of Epic's 100+ AI Use Cases Get the Most Adoption, What Changes About Physician Identity in the Age of AI, and more..." Click the play button to listen or read the show notes below.
‍

Audio:

‍

Guest(s):

• Jackie Gerhart, MD (@JackieGerhart), Chief Medical Officer at Epic Systems
• Joshua Liu, MD (@joshuapliu), Co-founder & CEO at SeamlessMD

‍

‍Episode 226 - Show Notes:

[00:00:07] Episode preview

[00:04:39] How two patient-care moments drove Dr. Gerhart from practicing medicine to joining Epic — After years as a family physician and OB in Madison, two incidents crystallized her shift: a sailboat trip where a friend pointed out she was "going in circles," and a breech delivery where a missed ultrasound led to a preventable C-section. Both made her ask how technology could fix systemic problems at scale, not just at the level of one practice.

[00:08:39] Why Epic prioritizes testing over interviews when hiring — Epic deliberately tests for curiosity, adaptability, grit, and resilience rather than credentials or interview polish. As founder and CEO Judy Faulkner frames it: if forced to choose, Epic prefers an inarticulate competent over an articulate incompetent. The same rigorous test battery applies across roles—including culinary staff on campus.

[00:10:39] What Dr. Gerhart's own Epic interview actually looked like — Applying post-call and caught off-guard by the requirement to have a mirror, she ended up sitting in her bathtub with her laptop because it was the only mirror she could find in her home. The tests span verbal, math, coding, and orthogonal-thinking problems—designed to measure speed, creativity, and reasoning process, not just correct answers.

[00:12:39] Why Epic consolidated all clinician-facing AI under the name "Art" — Rather than naming each of its 100-plus AI features individually, Epic rebranded them all under "Art," short for the art of medicine—not an acronym. The goal was to give clinicians and patients a coherent identity for what they're interacting with, and to let the features work together in concert rather than appear as disconnected tools.

[00:18:38] How Epic's ambient AI goes beyond note documentation — Chart with Art, now signed by 32 customers, was designed from the start around what Epic calls the Intelligent Visit. Rather than just capturing notes, it streams real-time decision support during the encounter—surfacing medication history, insurance coverage, and family history without the clinician needing to scroll or search. The goal is multimodal assistance across the entire patient journey, not just the office visit.

[00:22:38] The eliminate, automate, simplify framework for AI in clinical workflows — When asked how to prevent AI from becoming another source of pop-up fatigue, Dr. Gerhart described Epic's three-part principle: eliminate work that doesn't need a human (e.g., routing billing or scheduling messages away from the clinician's inbox entirely), automate repetitive clinical pathways, and simplify the data that remains so clinicians reach actionable insights faster.

[00:25:38] Which AI use cases are winning adoption—and why — Ambient documentation has driven the strongest clinician response, with multiple Chart with Art users telling Epic they're no longer considering retirement. Revenue cycle AI (billing, coding, prior authorizations) has also seen wide uptake due to its measurable, automatable nature. Chart summarization has been particularly well-adopted, with Epic finding roughly 25 percent of clinicians reviewing these summaries learn something new they wouldn't have otherwise found.

[00:27:38] How AI reading radiology reports is catching cancers earlier — Epic's AI reads the text of radiology reports—not the scans themselves—to flag incidental findings that might otherwise fall through the cracks. Christ Hospital implemented this and has identified multiple lung cancers at stage one that would likely have gone undetected until stage two, three, or four. Dr. Gerhart cited her own father's experience of a missed CT follow-up as a personal example of how these gaps happen even to clinically connected patients.

[00:30:38] How chart summarization is evolving toward agentic AI — Early versions involved prompt engineering and specialty-specific toggles for time range, verbosity, and specialty focus. The next phase runs multiple agents in sequence: for Chart with Art, one agent checks speech-to-text accuracy while a voice agent lets clinicians dictate their own note preferences into Haiku. The goal is specialty-level personalization driven by agent pipelines rather than individual prompt settings.

[00:32:38] Why AI adoption in medicine moves at the speed of trust — At a recent conference, Dr. Gerhart heard a line she found meaningful: "For physicians, AI moves at the speed of trust. For patients, it moves at the speed of desperation." Rather than describing an identity crisis, she frames physician uncertainty around AI as "identity curiosity"—a genuine rethinking of what unique value clinicians bring when knowledge itself becomes a commodity. Her answer: judgment, relationship, and the ability to meet a patient's specific needs in the moment.

[00:36:38] Why the human clinical touch still can't be replaced — Patients want someone caring who listens, understands their fear, and can explain options in language they understand. Dr. Gerhart's point: AI can deliver data and efficiency, but no AI sits on a hospital bed and holds a patient's hand through a stage-four cancer diagnosis—and most patients don't want it to. Clinicians who pair good judgment with AI become better, not redundant.

[00:38:38] How Epic named Emmy—and where the name "Epic" itself comes from — Emmy, the patient-facing AI in MyChart, was chosen for being approachable and warm without being falsely humanized. Its iconography is a simple heart. In explaining the naming process, Dr. Gerhart also revealed that "Epic" comes from the dictionary definition of Homer's epic—a glorious account of a nation's events—with "nation" swapped for "patient." Every product name goes through Judy Faulkner, who often names them herself.

[00:42:38] Why Emmy is better positioned for clinical questions than general-purpose AI — Unlike consumer tools, Emmy has the full context of a patient's clinical chart, operates as a HIPAA-covered entity under ONC standards, and keeps data within the EHR rather than sending it to an unregulated back end. Dr. Gerhart flagged the risk of patients using general-purpose AI for health questions—particularly the danger of not knowing what the AI gets wrong, which is most harmful when the person asking lacks the clinical training to catch the errors.

[00:48:38] How Epic thinks about competition with startups — Dr. Gerhart rejects the idea that the best standalone product always wins in healthcare IT, comparing it to using a smartphone camera over a professional camera for everyday life: integration into existing workflows matters more than isolated performance. Epic encourages customers to keep solutions that work and to give feedback on gaps—but notes that the real hurdle for startups is demonstrating deep workflow fit and change management support, not just product quality.

[00:52:38] What Epic is most misunderstood about—and how it responds to criticism — Three common misconceptions: why Epic remains private and builds everything in-house (to preserve a single interconnected codebase and avoid quarterly-earnings-driven tradeoffs); why it can't act as a "magic wand" across all customer deployments (each hospital configures its own instance); and why positive outcomes like MyChart often go unrecognized. For all feedback—positive or negative—Epic uses an internal framework called TEA: Thank, Evaluate, and Act. "This feedback that you're giving to me is a gift to me, and I need to actually do something to make it better."

‍

‍Fast 5 Lightning Round:

1. What is your favorite book or book you’ve gifted the most?
   The Innovator's Prescription by Clay Christensen
2. If you could instantly master any skill, what would it be?
   "I would wanna be like Jane Goodall and be able to speak animal... I guess I would try chimpanzee first... I feel like they bring a lot of joy to a lot of people's lives."
3. Would you rather have Super strength, super speed, or the ability to read people’s minds?
   Mind reading.
4. What is something in healthcare you believe others might find insane?
   "I love working on holidays... people don't come into urgent care unless they have some kind of problem that needs to be fixed that day... I feel like it's kind of this event, like we're all in this together."
5. What is the last movie or TV show you saw, and what did you think of it?
   "I did just reread All The Light You Cannot See, and saw that movie. I loved it... I just rewatched The Queen's Gambit. I really thought that was beautiful... I saw a musical with Hugh Jackman... it was about a magician, and he joined a circus... The Greatest Showman... That was good. It made me laugh and cry."

The Digital Patient has been recognized as Feedspot's #1 Patient Engagement Podcast of 2025. Thank you to our listeners for making this happen!