Abstract

Patients struggle with adhering to prescribed medication and leading a healthy lifestyle outside of clinical visits. A mechanism needs to be developed and deployed to constantly remind them about the need to take medication and to lead a healthy lifestyle in a non-intrusive way while watching for changes in vitals that can be collected using wearable devices. With these considerations, a Virtual Healthcare Assistant is proposed, built with a multi-agent system that is easily integrated with wearable devices such as Apple Watches, Fitbits, and Oura Rings. a LangGraph-based multi-agent workflow to mimic real-world continuous monitoring. The resulting system enables agent-to-agent communication, has low forecasting error, and proper anomaly detection. Overall, the architecture showcases interpretable, personalized, real-time health monitoring and provides a foundation for proactiveness in maintaining a healthy lifestyle and possible future clinical applications.

Recommended Citation

Mathew, Solomon T. and Mathew, Babu T. (2025) "Virtual AI Lifestyle Assistant," SMU Data Science Review: Vol. 9: No. 3, Article 6.
Available at: https://scholar.smu.edu/datasciencereview/vol9/iss3/6