SmartWalkCoach's AI agents boost walking motivation and UX in field study

SmartWalkCoach is a mobile AI companion developed by researchers (Xianzhe Zhang et al.) to support the full walking journey—from pre-walk planning to in-walk guidance and post-walk reflection. It bridges the gap between map navigation and motivational coaching by orchestrating three lightweight agents: GeographyAgent curates conversational routes using nearby points of interest and user preferences, offloading pathfinding to map APIs; AccompanyAgent delivers context-aware, just-in-time prompts that blend informational cues with relational encouragement; and SummaryAgent provides concise reflection and next-step planning. The end-to-end design aims to lower cognitive load during planning and sustain engagement through cadence-aware interventions.

In an in-the-wild, two-period AB/BA crossover study with 12 participants, the team compared Information-only guidance to Information+Motivation. Linear mixed models revealed that adding motivational, companion-like dialogue significantly improved positive feelings and user experience, with no evidence of carryover effects. Thematic analysis surfaced two design imperatives for mobile companions: supportive relational expression and context-aware timing (e.g., avoiding high-load moments, intervening at fatigue or milestones). The paper (to be presented at ACM IUI 2026) outlines limitations and paths toward multimodal, voice-first companions with adaptive personalization.