Designing a Fitness AI Coach Web Application (Evolve AI) for Personalized User Experience

The rapid proliferation of digital fitness applications has not been matched by a corresponding improvement in user retention or genuine personalization. Most fitness platforms offer static, one-size-fits-all workout plans that fail to account for individual goals, physical limitations, fitness levels, and behavioral patterns — resulting in high early abandonment rates. This research paper presents a comprehensive User-Centered Design (UCD) study conducted during the development of Evolve AI, an AI-powered fitness coaching web application designed to address this gap. The methodology encompasses structured user research (online survey: n=87; semi-structured interviews: n=12), competitive analysis of four major fitness platforms, user persona development, iterative wireframing, high-fidelity UI design using Figma, and two-round usability testing (n=24 participants). Special attention is given to the design of the AI recommendation user experience — specifically, how the interface communicates personalized workout suggestions in a transparent and trustworthy manner without technical complexity. Results demonstrate that targeted, evidence-based design improvements — specifically to the onboarding flow, AI recommendation interface, progress visualization, and navigation structure — resulted in a 23.3-point improvement in System Usability Scale (SUS) scores (from 68.4 to 84.7), a 34% increase in perceived personalization, and a 41% increase in reported intent to continue using the application. The paper further discusses design trade-offs, ethical considerations in AI-based health recommendations, real-world implementation challenges, and future directions including wearable integration, conversational AI coaching, and nutrition planning modules.