Abstract

This paper presents the iterative development of Habit Coach, a GPT-basedchatbot designed to support users in habit change through personalizedinteraction. Employing a user-centered design approach, we developed thechatbot using a Retrieval-Augmented Generation (RAG) system, which enablesbehavior personalization without retraining the underlying language model(GPT-4). The system leverages document retrieval and specialized prompts totailor interactions, drawing from Cognitive Behavioral Therapy (CBT) andnarrative therapy techniques. A key challenge in the development process wasthe difficulty of translating declarative knowledge into effective interactionbehaviors. In the initial phase, the chatbot was provided with declarativeknowledge about CBT via reference textbooks and high-level conversationalgoals. However, this approach resulted in imprecise and inefficient behavior,as the GPT model struggled to convert static information into dynamic andcontextually appropriate interactions. This highlighted the limitations ofrelying solely on declarative knowledge to guide chatbot behavior, particularlyin nuanced, therapeutic conversations. Over four iterations, we addressed thisissue by gradually transitioning towards procedural knowledge, refining thechatbot's interaction strategies, and improving its overall effectiveness. Inthe final evaluation, 5 participants engaged with the chatbot over fiveconsecutive days, receiving individualized CBT interventions. The Self-ReportHabit Index (SRHI) was used to measure habit strength before and after theintervention, revealing a reduction in habit strength post-intervention. Theseresults underscore the importance of procedural knowledge in driving effective,personalized behavior change support in RAG-based systems.