Mass General Brigham has begun using an AI intake agent embedded within MyChart, the patient portal used across much of the U.S. healthcare system, to handle primary care intake tasks before patients even see a clinician, part of a broader wave of hospital systems deploying AI agents to handle administrative work ahead of appointments.
What an AI Intake Agent Actually Does
Rather than patients filling out static forms or waiting for a nurse to collect basic information by phone, an AI intake agent can conduct an interactive, conversational pre-visit assessment inside the patient portal, gathering symptom details, medication history and visit reasons in a format that can be structured and fed directly into a clinician’s workflow before the appointment begins.
Why Primary Care Specifically Needs This
Primary care visits are often constrained to short appointment windows, and a significant portion of that limited time is traditionally spent on intake questions that could be gathered beforehand — making primary care one of the clearest use cases for AI agents that handle preparatory work, freeing clinician time for actual diagnosis and treatment discussion.
How This Connects to Epic’s Broader AI Push
The MyChart deployment arrives alongside Epic’s rollout of Connectors, which let its AI assistants retrieve information and complete tasks across other hospital software systems, reflecting a broader strategy among major electronic health record vendors to extend AI beyond simple chatbot interactions into agents that can actually complete multi-step administrative workflows inside the patient and clinician experience.
The Trust Question Hospitals Must Navigate
Patient intake agents raise a distinct trust challenge from diagnostic AI tools: patients are being asked to share sensitive health information with an AI system before any human has reviewed it, making transparency about how that information is used, stored and verified by a clinician afterward central to whether patients feel comfortable with the format rather than skipping or rushing through it.
What Skeptics of Rapid Rollout Point Out
Critics of fast AI intake deployment across large health systems note that intake data quality directly affects downstream clinical decisions, meaning errors or omissions introduced by an imperfect AI conversational agent could propagate into a patient’s record in ways a human intake nurse might have caught and corrected in the moment.
What’s Next
Watch for Mass General Brigham to publish data on patient completion rates and clinician satisfaction with the AI-gathered intake information, which will show whether the tool meaningfully improves visit efficiency or introduces friction patients and doctors find more cumbersome than the system it replaced.
Photo: Thirdman / PEXELS via Pexels
