Advocate Health and ECU Health have deployed autonomous AI agents on Epic Systems’ Agent Factory platform, demonstrating how health systems of different sizes can use no-code tools to reduce clinical workflow friction.
Advocate Health, which runs two of the largest Epic EHR instances globally, built agents for inpatient pharmacy verification and infusion chart preparation. The agents comb through patient histories, medications, and clinical notes to generate summaries pharmacists need for medication order verification. The health system’s existing internal AI and machine learning studio gave it a head start on development.
ECU Health, a smaller academic network in eastern North Carolina, worked directly with Epic developers on two prototypes: a transfer center agent that generates focused patient summaries and a discharge planning tool for case managers. The transfer agent saved roughly 20 hours of chart review time per week in its first month, the equivalent of half a full-time employee. The discharge tool collected mostly negative feedback on day one, but prompt refinements overnight pushed its approval rating to 75% by day two.
Both organizations say success with Agent Factory requires realistic expectations around token budgets, data governance, and clinician trust. Dr. John Lee, an Epic consultant, warned that organizations chasing AI out of FOMO will arrive unprepared. But for those with strong data foundations, the no-code platform enables custom autonomous workflows without middleware or external APIs. The deployments mark a practical step from AI copilots toward autonomous agents in frontline care.
