The anatomy of a personal health agent.

A comprehensive personal health agent integrating multimodal data analysis and personalized insights was developed and evaluated across 11 benchmark tasks, marking the most extensive assessment to date.

• Why it matters: Addressing diverse individual health needs remains a challenge in health support systems, highlighting the importance of personalized, accessible health agents that can interpret complex data and provide tailored guidance.
• What they did: The authors created a multi-agent framework comprising three specialist sub-agents—data science, health domain expert, and health coach—guided by user journeys identified through user and expert research, and evaluated it with over 7,000 annotations and 1,350 hours of expert and user effort.
• The result: This system's comprehensive evaluation demonstrates its potential to improve access to high-quality, personalized health support, laying a foundation for broader application and development in health technology.