Mass General Brigham, one of the largest hospital systems in New England, has rolled out an AI-powered intake agent inside its MyChart patient portal designed to help patients navigate primary care — from understanding which type of appointment they need to answering basic pre-visit questions before they see a clinician.
What the Intake Agent Does
The AI agent sits within the familiar MyChart interface patients already use for messaging providers and viewing test results, aiming to reduce the friction many patients experience when trying to figure out whether a symptom warrants a primary care visit, urgent care, or a specialist referral. By guiding patients through an initial triage-style conversation, the system is designed to route them more efficiently while also collecting structured information that can speed up the actual clinical encounter.
Addressing a Real Access Problem
Primary care access has become an increasingly acute problem across the U.S. healthcare system, with many patients facing weeks-long waits for appointments even as the physician workforce struggles to keep pace with demand. An AI intake agent won’t create more physician capacity, but it can theoretically reduce wasted visits — cases where a patient books an unnecessary primary care appointment for something better suited to urgent care, or conversely delays care because they’re unsure where to go.
Trust and Adoption Challenges
Patient portal AI tools face a trust hurdle that purely clinician-facing tools don’t: patients interacting with an AI agent about their own health concerns may be more anxious about errors, more likely to test the system with edge cases, or simply reluctant to trust a chatbot with sensitive symptom descriptions. Mass General Brigham has not published patient satisfaction or accuracy data from the rollout, making it difficult to assess how well the intake agent is actually performing in real-world use versus controlled testing.
A Broader EHR-Embedded AI Trend
The MyChart deployment fits a pattern seen elsewhere this year, including Northwestern Medicine’s integration of the OpenEvidence clinical evidence tool directly into Epic for physicians — both reflecting a shift toward embedding AI assistance natively within the software patients and clinicians already use daily, rather than requiring adoption of standalone AI apps. Epic, which underlies both MyChart and much of the EHR infrastructure at major health systems, has become a key battleground for where and how AI assistance gets surfaced.
Equity Concerns
Health equity researchers have raised questions about whether AI intake agents work equally well across different patient populations, including those with limited English proficiency, lower digital literacy, or complex, multi-symptom presentations that don’t fit neatly into structured triage categories. Mass General Brigham has not detailed what language support or accessibility accommodations the intake agent includes.
What’s Next
Mass General Brigham has not announced plans to expand the intake agent beyond primary care into other service lines, nor published outcome metrics such as changes in appointment no-show rates or emergency department diversion. Other large health systems are likely watching the rollout closely as a test case for whether AI-assisted patient intake can meaningfully ease access bottlenecks without introducing new equity or accuracy concerns.
Photo: Ivan S / PEXELS via Pexels
