GPTCoach explores how large language models can bridge the gap between generic mobile health apps and expensive human coaching services.

• Combines insights from 12 health professionals and 10 potential users to develop LLM-based coaching
• Implements evidence-based coaching principles in a scalable digital format
• Creates personalized activity plans that adapt to individual needs and contexts
• Demonstrates how AI can democratize access to quality health coaching

This innovation matters because it addresses healthcare accessibility challenges by offering personalized support at scale, potentially improving physical activity outcomes for diverse populations.

GPTCoach: Towards LLM-Based Physical Activity Coaching