01. Project Scope
Harmona is an AI-powered wearable companion designed to help women athletes improve recovery, stabilize energy, and reduce training-related burnout. The goal was to explore how multimodal data (HRV, sleep, menstrual cycle, mood, and activity load) could be translated into clear, supportive, and trustworthy guidance.
My responsibility was to design the core interaction model and create a white-box explanation layer (XAI) to help users understand why Harmona recommends specific recovery actions.
Objectives:
- Define the user problem and opportunity space
- Map the wearable data inputs and AI decision pipeline
- Create a transparent explanation model (XAI) to increase user trust
- Design the UI patterns for daily insights, nudges, and recommendations
- Prototype a mobile app experience that feels empathetic, feminine, and athletic
02. Process
Research & Problem Framing
Women athletes often experience:
- Energy crashes across the hormonal cycle
- Difficulty interpreting HRV/sleep metrics
- Training inconsistency due to poor recovery timing
- Low trust in generic, “black box” wearables
I reviewed academic research, wearable UX patterns, women’s health cycle models, and recovery scoring methods to ground the concept in evidence-based frameworks.
Data Architecture & AI Pipeline
I created a high-level 4-stage model that defined how Harmona’s insights would function:
- Input Data
HRV, sleep, mood, cycle phase, optional training load - AI Pattern Detection
ML identifies patterns across personal + anonymized cohorts - White-Box Explanation (XAI)
Clear reasoning behind each recommendation - Supportive Nudge Options
Actionable choices (e.g., breathwork, fueling, rest timing)
This framework balanced personalization with user control, avoiding overly prescriptive or paternalistic coaching.
Interaction Design
I designed:
- The daily guidance flow
- The insight → explanation → action loop
- A soft, calming visual language using gradient “energy states”
- Modular components for recovery cards, cycle insights, and AI explanations
- Optional voice and haptic interactions for hands-free moments
The UI is feminine without being cutesy — inspired by sport science, wellness calm, and high-end wearables.
Prototyping
Using Figma, I built:
- High-fidelity screens
- Animated micro-interactions
- A guided onboarding experience
- The XAI explanation drawer design
- A system of states for various recovery conditions
- Early prompt structures for an AI assistant
03. Solution
A Supportive, Intelligent Companion for Women Athletes
Harmona delivers:
- Daily Insight Cards that summarize recovery status
- Why This Matters explanations powered by transparent AI
- Personalized Nudges, like fueling tips or breathwork sessions
- Cycle-Aware Coaching that adapts to hormonal phases
- Adaptive Prompts for journaling, mood, or energy check-ins
The result is an experience that feels:
- Human (empathetic, encouraging)
- Scientifically grounded (XAI backing each insight)
- Actionable (simple, achievable micro-choices)
- Trust-building (no mystery “scores”)
Harmona positions itself as a recovery companion, not a tracker — a coach that helps women understand their bodies, protect their energy, and train smarter.
04. Key Takeaways
- XAI dramatically increases trust in health recommendations
- Women want choice, not prescriptions, supportive autonomy is crucial
- A feminine athletic aesthetic can feel premium and modern
- Wearables need context, not numbers
Harmona is a vision for the next chapter of women’s health tech — one that blends AI, empathy, and science.
