UX Case Study

PulseIQ

An AI-enhanced fitness and wellness app designed to help users build sustainable healthy habits through personalized workout prescriptions, adaptive coaching, and empathetic progress tracking.

Context
Designlab · AI for UX
Role
Senior Product Designer
Methods
Discovery · Research · Prototyping · Testing
Tools
Claude Pro · ChatGPT · Gemini · Firefly
Type
Mobile App · B2C


PulseIQ

Discovery

The Brief

PulseIQ is an AI-enhanced fitness and wellness app designed to help users build sustainable healthy habits through personalized workout prescriptions and progress tracking. It combines smart onboarding with adaptive coaching to meet users at their current fitness level.

The central design challenge was to make AI-driven guidance feel supportive rather than prescriptive — translating complex personalization logic into moments that feel human, empathetic, and trustworthy.

"Help people train smarter, not harder — and never make them feel like they're starting from zero."

Brief

Brand Voice

Before any interface work began, I defined PulseIQ's brand voice as a design constraint — a north star for every content decision, from onboarding copy to AI coaching messages.

Brand Voice Principles
Confident and motivating — the clarity of a knowledgeable coach blended with the encouragement of a trusted training partner.
Intelligently human — AI-driven insights communicated in clear, approachable language. Never algorithmic, never cold.
Progress over perfection — energetic yet grounded. Motivates without hype. Celebrates consistency, not just peaks.
Personal and empathetic — adapts to each user's real-life constraints and goals. Designed for actual humans, not aspirational avatars.
Brand Voice

Visual Identity & Moodboard

PulseIQ's visual identity was designed to support clarity, motivation, and everyday usability. A green-forward palette, rounded components, and soft depth were chosen to create a calm, approachable environment — deliberately avoiding the dark-neon aesthetic that dominates the fitness app space.

Brand elements and UI patterns were developed alongside user goals: reducing intimidation, reinforcing progress, and making guidance feel supportive rather than prescriptive. Modular cards and data visualizations prioritize scannability and quick comprehension.

Logo
Moodboard

Brand System

This brand system documents the complete visual and verbal identity for PulseIQ. The work covers foundations through to interface — color palette with tint ramps and semantic tokens, a nine-level type scale, logo usage rules with clear space and minimum size guidance, brand voice principles with channel-specific tone guidance and a do/don't vocabulary guide, iconography, photography direction, a full UI component library, and a spacing and grid reference. The system was built to serve both design and development handoff: every section is precise enough to act as a production reference, while the interactive HTML format keeps it accessible across devices. The design decisions throughout reflect the core brand tension — clinical enough to be trusted by a health-conscious audience, warm enough to feel like a coach rather than an app.

Brand System

User Personas

Three personas represent health-motivated adults aged 40–60 — a segment systematically underserved by mainstream fitness apps. Each maps to a distinct entry point into PulseIQ's onboarding pathway, with specific movement restrictions, clinical context, and motivational profile.

RB
Robert, 57
Operations manager · Doctor-referred
Elevated blood pressure, rising LDL, pre-diabetic reading. Physically capable and motivated by prevention — not rehabilitation. Needs to know what intensity is safe for his cardiovascular profile, and something to show at his three-month follow-up.
heart-rate-awaremedical-referralprevention-focusedshort-session
DW
Diane, 61
Retired nurse · Post-surgical
Hip replacement six months ago. Cleared by her physical therapist with specific restrictions: no high-impact movement, no deep hip flexion, no twisting. Every app she's tried either ignores her limits or treats her like she's still a patient.
low-impactgentle-startmedical-referralpost-surgical
CM
Carmen, 42
Program Manager · Chronic back pain
Herniated disc at L4-L5 and chronic stress — both self-managed, both connected. She knows exactly what sets her back off. She needs an app that respects what she already knows about her body and adapts when her energy and pain levels vary.
low-impactstress-firstenergy-variablejoints-back-sensitive

Market Research

I leveraged AI to analyze competitive landscapes and synthesize user research — uncovering unmet needs and stress-testing whether AI-generated outputs met inclusive design standards. The goal was to identify gaps that PulseIQ could own, not just match the market.

User Journey Map

A representation of the user's end-to-end experience — actions, emotions, and mental models at each stage. Mapped across Carmen's journey from initial awareness through the critical retention moment of returning after a missed workout.

Two key pain points surfaced from this mapping that became central design constraints:

Pain Point 1 — The Return Moment

Post-missed-day re-entry is a retention killer. Fear of "starting over" is strong. Micro-copy at re-entry isn't cosmetic — it's a conversion moment. The interface must communicate flexibility without abandoning the user.

Pain Point 2 — Trust Barrier

AI recommendations feel arbitrary without visible reasoning. Surfacing the "why" behind suggestions isn't just about transparency — it's the mechanism by which users build long-term trust in the system.

Journey Map
Journey map — Carmen, 42. Awareness through Month 1 retention.

Market & Feature Gap Analysis

A structured evaluation comparing user needs and competitor offerings — identifying missing or underdeveloped features across the fitness app landscape. This analysis directly informed PulseIQ's differentiation strategy and feature prioritization.

Feature / Need Zing Coach Freeletics Fitbod PulseIQ Opportunity
Adaptive AI coaching (personalized to real-time feedback)PartialPartialPartialFull adaptive loop — plans evolve with every session and health signal
Health condition-aware workout modificationPartialGapGapOnboarding captures health context; AI adjusts safely without requiring diagnoses
Non-judgmental re-entry after missed sessionsGapGapGapDesigned as a core experience moment — warm, low-pressure, and built to bring people back
Transparent AI reasoning ("why this workout")GapGapGapExplanations surfaced inline; users can understand and trust every recommendation
Integrated nutrition and fitness in one experienceGapPartialGapNutrition and movement connected in one view — no separate apps or subscriptions
Mental wellness and stress-aware programmingGapPartialGapMood and energy check-ins shape daily recommendations — the whole person, not just the workout

Prototyping & Testing

I used Claude Pro to create interactive, high-fidelity representations of the product — testing ideas, gathering feedback, and validating design decisions before any development investment. The prototype covers the full user journey across five core flows.

Screen Flow Diagram

Before building high-fidelity screens, I mapped the complete architecture: all screens, decision points, and navigation paths. This diagram served as the single source of truth for scope — preventing scope creep and aligning stakeholder expectations before design work began.

01
Onboarding
Goals, fitness level, injury history, schedule. 7 screens.
02
Dashboard
Today's plan, progress ring, recent activity summary.
03
Workout
AI prescription, exercise detail, in-session coaching.
04
Nutrition
Meal log, macro overview, AI nutrition insights.
05
Profile
Progress history, settings, connected devices.
Screen flow
Screen flow diagram — complete navigation architecture across all 5 user flows. View full diagram ↗

Prototype Highlights

The high-fidelity prototype brings PulseIQ's core user journeys to life in a clickable, phone-framed simulation. It serves as a design and communication tool — aligning on screen architecture, copy, and AI interaction patterns before any development begins.

5
core user flows prototyped
3
personas validated against
AI
built with Claude Pro
100%
mobile-native interaction patterns
Onboarding That Listens

Smart onboarding captures fitness level, goals, injury history, and schedule constraints. Users feel understood, not processed. No generic questionnaire vibes.

Adaptive AI Coaching

After each session, the app asks "How did that feel?" — not a rating, a check-in. AI adjusts the next workout based on real feedback, not just completion data.

Non-Judgmental Re-entry

Missing days doesn't reset progress. The app adjusts without guilt — treating gaps as data, not failure. Micro-copy at re-entry is deliberate and tested.

Transparent AI Reasoning

Every AI recommendation surfaces its rationale inline. Users can see why a workout was prescribed — building trust gradually rather than demanding it upfront.

Modular Data Cards

Progress shown through consistency metrics, not just calories burned or streaks. The dashboard prioritizes the story of the week, not the single-session highlight.

Inclusive by Design

All AI-generated content was audited against the bias framework developed in Phase 1. Tone, imagery references, and metric framing were corrected for diverse users.

Prototype
Try the interactive prototype
A clickable, phone-framed simulation covering all 5 core flows — built with Claude Pro.
Try the prototype

Key Learnings

What AI Changed

Using AI throughout the design process — for competitive research, persona synthesis, moodboard generation, and prototype building — dramatically compressed the time between question and answer. But the most important skill was knowing when to override.

The AI's default direction for a fitness app was predictably "Tech Coach" — dark UI, neon green, high-energy sport photography. That direction was rejected in the first review because it defaulted to a young male athlete aesthetic that excluded the core personas. The brief explicitly called for "intelligently human" — and that required human judgment to define.

"AI accelerates the distance between idea and artifact. The designer's job is still to decide which artifacts are worth making."

What Didn't Change

Empathy can't be automated. The bias audit, the decision to reframe "missed days" as data rather than failure, the choice to surface AI reasoning rather than hide it — these were human design decisions grounded in the personas, not in AI-generated defaults.

The three things that AI couldn't do for this project:

Define the emotional north star

Progress over perfection. No judgment. These were values — not features — that shaped every subsequent decision.

Run the bias audit

AI surfaced assumptions. A designer had to decide which ones mattered and how to correct them without introducing new ones.

Make the re-entry moment feel human

The micro-copy for returning after a missed workout — that required understanding the psychology of habit formation, not just UX patterns.

← Previous
Next →