Case Study

Aira Health

Aira Health

Support for shift workers who work irregular schdules

Support for shift workers who work irregular schdules
Overview
Roles

Researcher, Interaction Design, Visual Design, Wireframing, Prototyping

Timeline

8 months

Tools

Figma, Illustrator

Health reminders regarding your health and how to stay on top of it.

Aira is an AI-powered health and wellbeing app designed for people whose lives do not follow predictable 9-5 routines, meant help make sense of daily health signals without pressure, guilt, or rigid optimization frameworks. It translates fragmented inputs — hydration, stress, mood, and sleep — into adaptive, emotionally intelligent guidance.

The goal was not to build another calorie counter or step tracker, but to reimagine health technology as a flexible companion that supports irregularity rather than punishing it. The core premise is to: translate fragmented health data (hydration, stress, mood, sleep) into humane, supportive guidance that adapts to real life.

Overview
Roles

Researcher, Interaction Design, Visual Design, Wireframing, Prototyping

Timeline

8 months

Tools

Figma, Illustrator

Health reminders regarding your health and how to stay on top of it.

Aira is an AI-powered health and wellbeing app designed for people whose lives do not follow predictable 9-5 routines, meant help make sense of daily health signals without pressure, guilt, or rigid optimization frameworks. It translates fragmented inputs — hydration, stress, mood, and sleep — into adaptive, emotionally intelligent guidance.

The goal was not to build another calorie counter or step tracker, but to reimagine health technology as a flexible companion that supports irregularity rather than punishing it. The core premise is to: translate fragmented health data (hydration, stress, mood, sleep) into humane, supportive guidance that adapts to real life.

Context

Health tracking today is built on the assumption of regularity: eight hours of nightly sleep, meals at consistent times, and activity that neatly fits into daily routines. But for millions of people, this isn’t reality. Nurses rotate shifts, ride share drivers stay up all night, parents wake every few hours, and students live in cycles of exams and deadlines. The consequences of these irregular lifestyles are disrupted circadian rhythms, chronic fatigue, increased risk of cardiovascular disease, and heightened stress.

Yet most health trackers are built around consistency, punishing users with “broken streaks” when life becomes unpredictable. In contrast, Aira prioritizes adaptability over rigidity and empathy over judgment, meeting users where they are rather than where a system expects them to be.

Guilt-Inducing Streaks

Track everything consistently. When users can’t keep up, the product response is often reminders, warnings, or silence, all of which quietly signal failure.

Inaccurate Predictions or Manual Input

Systems fail to account for daytime sleep or fluctuating energy cycles.

The "Failure" Signal

When a user cannot meet a benchmark, the app responds with warnings or silence, signaling failure

Context

Health tracking today is built on the assumption of regularity: eight hours of nightly sleep, meals at consistent times, and activity that neatly fits into daily routines. But for millions of people, this isn’t reality. Nurses rotate shifts, ride share drivers stay up all night, parents wake every few hours, and students live in cycles of exams and deadlines. The consequences of these irregular lifestyles are disrupted circadian rhythms, chronic fatigue, increased risk of cardiovascular disease, and heightened stress.

Yet most health trackers are built around consistency, punishing users with “broken streaks” when life becomes unpredictable. In contrast, Aira prioritizes adaptability over rigidity and empathy over judgment, meeting users where they are rather than where a system expects them to be.

Guilt-Inducing Streaks

Track everything consistently. When users can’t keep up, the product response is often reminders, warnings, or silence, all of which quietly signal failure.

Inaccurate Predictions or Manual Input

Systems fail to account for daytime sleep or fluctuating energy cycles.

The "Failure" Signal

When a user cannot meet a benchmark, the app responds with warnings or silence, signaling failure

Problem Statement

Standard health technology is built on the "Optimization Trap", the assumption of 8 hours of sleep, regular meals, and steady activity. For shift workers, students, and new parents, these apps become a source of guilt rather than support. The design opportunity is to create a health tracker that uses AI to adapt to unpredictability, providing flexible, personalized guidance that empowers users rather than judging them.

Solution

Aira is an AI-powered health tracker built specifically for irregular lives. It uses adaptive pattern recognition to learn each user's unique rhythms, predicting fatigue and delivering empathetic, actionable guidance.

Solution

Aira is an AI-powered health tracker built specifically for irregular lives. It uses adaptive pattern recognition to learn each user's unique rhythms, predicting fatigue and delivering empathetic, actionable guidance.

Research

User Research

Personas

I mapped personas across different irregular lifestyles. For each of these users, the core use cases include sleep recovery guidance, fatigue alerts, flexible goal setting, and wellness tracking that aids messy, unpredictable schedules instead of penalizing them.

Personas

I mapped personas across different irregular lifestyles. For each of these users, the core use cases include sleep recovery guidance, fatigue alerts, flexible goal setting, and wellness tracking that aids messy, unpredictable schedules instead of penalizing them.

Competitive Analysis

Mainstream Trackers (Apple Health, Fitbit): Descriptive but assume regularity; they offer irrelevant insights for non-traditional schedules.

High-Performance Tools (Oura, Whoop): Use predictive analytics but target elite athletes and "biohackers" rather than everyday users managing chaos.

Scheduling Apps: Help organize time but offer no guidance on the health impacts of those schedules.

Competitive Analysis

Mainstream Trackers (Apple Health, Fitbit): Descriptive but assume regularity; they offer irrelevant insights for non-traditional schedules.

High-Performance Tools (Oura, Whoop): Use predictive analytics but target elite athletes and "biohackers" rather than everyday users managing chaos.

Scheduling Apps: Help organize time but offer no guidance on the health impacts of those schedules.

Design Process

Design Goals & Principles

The system uses predictive AI to adapt to real-life variability, anticipating fatigue through a supportive and non-judgmental lens. This flexible framework ensures inclusivity by reframing setbacks as natural adaptations rather than failures.

  1. Human over system

    • The app should speak like a supportive companion, not a dashboard or protocol.

  2. Adaptation over discipline

    • Plans should flex based on context, not punish inconsistency.

  3. Interpretation over raw data

    • Users don’t need more numbers; they need meaning.

  4. Permission to pause

    • Rest, switching, or disengaging should be normalized.

These principles became especially important later when system complexity increased.

Design Goals & Principles

The system uses predictive AI to adapt to real-life variability, anticipating fatigue through a supportive and non-judgmental lens. This flexible framework ensures inclusivity by reframing setbacks as natural adaptations rather than failures.

  1. Human over system

    • The app should speak like a supportive companion, not a dashboard or protocol.

  2. Adaptation over discipline

    • Plans should flex based on context, not punish inconsistency.

  3. Interpretation over raw data

    • Users don’t need more numbers; they need meaning.

  4. Permission to pause

    • Rest, switching, or disengaging should be normalized.

These principles became especially important later when system complexity increased.

Considerations

I moved the AI’s role from simple data reporting to proactive adaptation through several core functions:

  • Adaptive Pattern Recognition: Learning unique, irregular rhythms over time.

  • Scenario Simulation: Previewing "what-if" health strategies for upcoming shifts or schedules.

  • Fatigue Alerts: Using wearable data to provide early warnings before exhaustion sets in.

  • Actionable Nudges: Delivering personalized, encouraging recommendations for immediate needs.

  • Empathetic Framing: Replacing "failure" metrics with supportive, recovery-focused language.

Considerations

I moved the AI’s role from simple data reporting to proactive adaptation through several core functions:

  • Adaptive Pattern Recognition: Learning unique, irregular rhythms over time.

  • Scenario Simulation: Previewing "what-if" health strategies for upcoming shifts or schedules.

  • Fatigue Alerts: Using wearable data to provide early warnings before exhaustion sets in.

  • Actionable Nudges: Delivering personalized, encouraging recommendations for immediate needs.

  • Empathetic Framing: Replacing "failure" metrics with supportive, recovery-focused language.

Information Architecture & User Flow

Given the user's irregular schedule, the hierarchy focuses on surfacing the most critical information with immediate accessibility.

The app is organized into four distinct tabs:

  • Home (The Dashboard): Acts as the central hub for daily check-ins, immediate summaries, and emotional reassurance.

  • Insights: A space for data-driven reflections. Instead of raw charts, it uses "Insight Cards" to show patterns and predictive guidance (e.g., "Plan a 20-min nap around 5pm").

  • Plans: The engine of the app. It houses guided support paths like Fatigue Management or Post-Shift Recovery, prioritized by Smart Discovery.

  • Profile: Centralizes personal settings, work rhythm preferences, and wearable data integrations.

Information Architecture & User Flow

Given the user's irregular schedule, the hierarchy focuses on surfacing the most critical information with immediate accessibility.

The app is organized into four distinct tabs:

  • Home (The Dashboard): Acts as the central hub for daily check-ins, immediate summaries, and emotional reassurance.

  • Insights: A space for data-driven reflections. Instead of raw charts, it uses "Insight Cards" to show patterns and predictive guidance (e.g., "Plan a 20-min nap around 5pm").

  • Plans: The engine of the app. It houses guided support paths like Fatigue Management or Post-Shift Recovery, prioritized by Smart Discovery.

  • Profile: Centralizes personal settings, work rhythm preferences, and wearable data integrations.

User Flow

This flow transitions from proactive onboarding to real-time support, ensuring the AI anticipates needs rather than just reacting to data.

1. Onboarding (Baseline)

A data-driven setup that maps the user's specific wellness goals, shift rhythms (fixed or rotating), and current state to immediately calibrate the AI's logic.

2. The Daily Loop (Contextual Support)

  • Pre-Shift: Predictive meal and nap strategies.

  • Intra-Shift: Real-time fatigue alerts triggered by wearable data.

  • Post-Shift: Personalized recovery coaching and progress re-framing.

3. Weekly Reflection (Long-term Growth)

A low-pressure review that identifies performance patterns and suggests non-judgmental adjustments for the week ahead.

User Flow

This flow transitions from proactive onboarding to real-time support, ensuring the AI anticipates needs rather than just reacting to data.

1. Onboarding (Baseline)

A data-driven setup that maps the user's specific wellness goals, shift rhythms (fixed or rotating), and current state to immediately calibrate the AI's logic.

2. The Daily Loop (Contextual Support)

  • Pre-Shift: Predictive meal and nap strategies.

  • Intra-Shift: Real-time fatigue alerts triggered by wearable data.

  • Post-Shift: Personalized recovery coaching and progress re-framing.

3. Weekly Reflection (Long-term Growth)

A low-pressure review that identifies performance patterns and suggests non-judgmental adjustments for the week ahead.

Usability Testing

Initial testing with users, particularly nurses, provided critical insights:

  • Proactive vs. Reactive: Users emphasized the need for insights before fatigue set in, not after they were already exhausted.

  • Reduced Cognitive Load: I simplified AI explanations to ensure they felt intuitive during high-stress work blocks.

Usability Testing

Initial testing with users, particularly nurses, provided critical insights:

  • Proactive vs. Reactive: Users emphasized the need for insights before fatigue set in, not after they were already exhausted.

  • Reduced Cognitive Load: I simplified AI explanations to ensure they felt intuitive during high-stress work blocks.

Challenges

  1. Overall Health Plans

The Overall Health category became a microcosm of the entire project’s complexity. Initially, the volume of content made it difficult to maintain a distinct identity for each plan without them feeling redundant or clinical.

  • From Categories to Concerns: I shifted the nomenclature from rigid "Health Categories" to user-focused "Health Concerns," making the plans feel more immediate and relevant to daily life.

  • The Triad Structure: To manage the volume, I established a consistent structure for every category: Steady, Fluctuation, and Strain. This allowed for three distinct plans per category (e.g., Emotional Awareness, Mood Stability, Daily Decompression), ensuring the system could support a user in any state.

  • Cohesion without Overlap: Each individual plan was reworked to focus on "rhythms over rules," ensuring they remained cohesive as part of the AIRA ecosystem while maintaining clear boundaries to prevent overlapping guidance.

  1. Overall Health Plans

The Overall Health category became a microcosm of the entire project’s complexity. Initially, the volume of content made it difficult to maintain a distinct identity for each plan without them feeling redundant or clinical.

  • From Categories to Concerns: I shifted the nomenclature from rigid "Health Categories" to user-focused "Health Concerns," making the plans feel more immediate and relevant to daily life.

  • The Triad Structure: To manage the volume, I established a consistent structure for every category: Steady, Fluctuation, and Strain. This allowed for three distinct plans per category (e.g., Emotional Awareness, Mood Stability, Daily Decompression), ensuring the system could support a user in any state.

  • Cohesion without Overlap: Each individual plan was reworked to focus on "rhythms over rules," ensuring they remained cohesive as part of the AIRA ecosystem while maintaining clear boundaries to prevent overlapping guidance.

  1. The Choice Fatigue Challenge

Problem: Originally, users had to manually navigate a "Masterlist" to find plans. Testing revealed this was overwhelming for tired shift workers.

The Solution: I pivoted to Smart Discovery, where the AI performs the heavy lifting by recommending plans based on real-time health signals. This shifted AIRA from being a "manual tool" to a "responsive companion".

  1. The Choice Fatigue Challenge

Problem: Originally, users had to manually navigate a "Masterlist" to find plans. Testing revealed this was overwhelming for tired shift workers.

The Solution: I pivoted to Smart Discovery, where the AI performs the heavy lifting by recommending plans based on real-time health signals. This shifted AIRA from being a "manual tool" to a "responsive companion".

Final

Reflection
Healthcare + Design

Originally, I chose this topic and specified it because of my lived experience as a healthcare worker and being on an irregular schedule compared to my friends and family. I felt that this was a population that was overlooked often because they were essential workers and I wanted to translate my experience into a digital solution to address real-world fatigue and recovery needs. This project allowed me to utilize and combine both fields of study and grow as a professional in both regards.

Less Equals Calm

I learned that "calm is a feature" and that the most difficult part of AI design is knowing when to hold back. Moving away from over-engineered, prescriptive logic, I prioritized providing users with meaning and agency over raw data.

Redefining Success Metrics

In healthcare and design, success is often measured by rigid compliance, but this project taught me to value a system that is adaptable to individual needs. Which is fundamental to holistic design and user experience solutions.

Reflection
Healthcare + Design

Originally, I chose this topic and specified it because of my lived experience as a healthcare worker and being on an irregular schedule compared to my friends and family. I felt that this was a population that was overlooked often because they were essential workers and I wanted to translate my experience into a digital solution to address real-world fatigue and recovery needs. This project allowed me to utilize and combine both fields of study and grow as a professional in both regards.

Less Equals Calm

I learned that "calm is a feature" and that the most difficult part of AI design is knowing when to hold back. Moving away from over-engineered, prescriptive logic, I prioritized providing users with meaning and agency over raw data.

Redefining Success Metrics

In healthcare and design, success is often measured by rigid compliance, but this project taught me to value a system that is adaptable to individual needs. Which is fundamental to holistic design and user experience solutions.

What's Next

Plans Adjustment: Based on the user's interactions with the plans and their retention I would re-evaluate if the plans I created are appropriate and which ones user prefer.

Long-term Insight Loops: I will shift the focus from daily performance to long-term trends. By helping users reflect on patterns over weeks and months, the app reinforces self-awareness and effort instead of relying on the pressure of daily consistency.

What's Next

Plans Adjustment: Based on the user's interactions with the plans and their retention I would re-evaluate if the plans I created are appropriate and which ones user prefer.

Long-term Insight Loops: I will shift the focus from daily performance to long-term trends. By helping users reflect on patterns over weeks and months, the app reinforces self-awareness and effort instead of relying on the pressure of daily consistency.