Talking to Your Data: Exploring Embodied Conversation as an Interface for Personal Health Reflection

The article presents a novel interaction paradigm for personal health data analysis using an embodied conversational agent, which integrates lightweight data preprocessing with a Unity-based character. The system features a dual-agent design, where an Observer agent extracts statistics and trends, while a Presenter agent communicates these findings without offering clinical advice. This approach was evaluated through a user study comparing traditional dashboards with conversational reflection, highlighting the potential for improved user engagement and understanding in health data interpretation, which is crucial for developing more intuitive AI interfaces in health technology.