CASE STUDIES
Amazon Halo
Amazon Halo
Amazon Halo
Chasing product-market fit
Chasing product-market fit
MY ROLE
Design Lead
USERS
Fitness Lapsed
BUSINESS GOAL
New Amazon Flywheel
CHALLENGE
Product-Market Fit
Nascent multimodal fitness ecosystem
Nascent multimodal fitness ecosystem
Unique but disconnected AI-powered features
Discover
Activity
Sleep
Nutrition
AI-POWERED FEATURES
Body Composition
Movement
Tone analysis
Unique but disconnected AI-powered features
Discover
Activity
Sleep
Nutrition
AI-POWERED FEATURES
Body Composition
Movement
Tone analysis
Unique but disconnected AI-powered features
Discover
Activity
Sleep
Nutrition
AI-POWERED FEATURES
Body Composition
Movement
Tone analysis
The launch.
The previously delayed v.1 launch is behind schedule
Feed – at launch
“Hub” – at launch
GOAL
Unblock engineering
ACTION
Prioritize and finish approved designs for launch and fast follows.
OUTCOME
Made the launch date!
My start.
The previously delayed v.1 launch is behind schedule
Feed – at launch
“Hub” – at launch
GOAL
Unblock engineering
ACTION
Prioritize and finish approved designs for launch and fast follows.
OUTCOME
Made the launch date!
GOAL
Make the Feed and Hub feel personal and helpful.
ACTION
Perform user research to inform new Feed content categories, usability improvements, and the ability to see Hub metrics at a glance.
OUTCOME
CSAT scores went up and churn rates went down.
Post-launch.
Personalization not working, usability hurdles, adding moments of delight
Personalization not working, usability hurdles, adding moments of delight
Feed iteration – morning
Hub iteration – at-a-glance
App
Highlights
Milestones
& Insights
Feedback
Nudges
New Feed – morning
New Feed – new content
New Feed – afternoon + milestone
New Hub – at-a-glance
Breakout new feature?
We used computer vision to create workouts that could see and react to your form.
Mirror/cast to TV, then prop up the phone on the floor with the camera angled correctly
Coach and video react to your form, tempo, and positioning in frame
Coach and video react to your form, tempo, and positioning in frame
See finally results on phone
See finally results on phone
Celebration, highlights, and workout insights details
GOAL
Test whether this could become a core Halo experience.
Action
Stepped in to design the interactive workouts and the companion FireTV app.
Outcome
Interactive workouts became one of our most popular workout types.
Goal
Create an innovative, sustainable, buildable product
Action
Organized a cross-functional design sprint to reimagine the entire end-to-end journey
Outcomes
All three concepts were combined into our new north star
Sign ups still flat.
People don't need more inspiration. They need a personalized plan, accountability, and help overcoming barriers.
Three strong concepts emerged
1
AI-powered
Personal fitness plan
2
AI-powered
Progress Tracking
3
Conversational UI
New North Star.
An AI-powered personal trainer in your pocket
HIGHLIGHT
New Visual Direction
Partnered with the brand team as the first step in building a new design system.
HIGHLIGHT
Personalized
Fitness Plans
The plan is a commitment, but also supportive and flexible. It anticipates blockers and helps users build resilience as they work through them.
AI features center stage.
Novel AI features are now a first step in building a personalized fitness plan
Key differentiator.
Our v.1 interactive workouts feature now core to the new experience
HIGHLIGHTS
Baseline interactive workouts
First workout calculates baseline fitness and form to inform workout progression
Checkin interactive workouts
Checkins track progress against the baselines. Makes adjustments as necessary on future workouts. Help keep motivation up.
HIGHLIGHT
Personalized
Fitness Plans
The plan is a commitment, but also supportive and flexible. It anticipates blockers and helps users build resilience as they work through them.
12 weeks of work!
A celebration, final fitness report, and nudge for the next round
What happened?
Going from evolution to revolution
v1
Increased CSAT, reduced churn
With personalization broken, designed a process to create ongoing, useful, fun, low-lift Feed content to address users' needs and frustrations
v2
Aligned org on building a redesigned, refocused product
The AI Trainer concept and beta testing had very enthusiastic feedback.
Increased app engagement
Interactive workouts became one of our most popular workout types.
Profitability time horizon proved too much for leadership
Andy Jassy became CEO and brought a more cold-eyed assessment to big bets like this one. He sunsetted the product.
CASE STUDIES
Amazon Halo
Chasing product-market fit
MY ROLE
Design Lead
USERS
Fitness Lapsed
BUSINESS GOAL
New Amazon Flywheel
CHALLENGE
Product-Market Fit
Nascent multimodal fitness ecosystem
Unique but disconnected AI-powered features
Discover
Activity
Sleep
Nutrition
AI-POWERED FEATURES
Body Composition
Movement
Tone analysis
My start.
The previously delayed v.1 launch is behind schedule
GOAL
Unblock engineering
ACTION
Prioritize and finish approved designs for launch and fast follows.
OUTCOME
Made the launch date!
Feed – at launch
“Hub” – at launch
Post-launch iterations.
Personalization not working, usability hurdles, adding moments of delight
GOAL
Make the Feed feel personal and helpful. Allow users to see their "Hub" data at a glance.
ACTION
Create and populate new Feed content categories. Use animation as more than just decoration.
OUTCOME
CSAT scores went up and churn rates went down.
Feed iteration – morning
Hub iteration – at-a-glance
App
Highlights
Milestones
& Insights
Feedback
Nudges
Breakout new feature?
We used computer vision to create workouts that could see and react to your form.
GOAL
Test whether this could become a core Halo experience.
Action
Stepped in to design the interactive workouts and the companion FireTV app.
Outcome
Interactive workouts became one of our most popular workout types.
Mirror/cast to TV, then prop up the phone on the floor with the camera angled correctly
Coach and video react to your form, tempo, and positioning in frame
See finally results on phone
Sign ups still flat.
People don't need more inspiration. They need a personalized plan, accountability, and help overcoming barriers.
Goal
Create an innovative, sustainable, buildable product
Action
Organized a cross-functional design sprint to reimagine the entire end-to-end journey
Outcomes
All three concepts were combined into our new north star
Three strong concepts emerged
1
AI-powered
Personal fitness plan
2
AI-powered
Progress Tracking
3
Conversational UI
The AI Trainer is born
An AI-powered personal trainer in your pocket
HIGHLIGHT
New Visual Direction
Partnered with the brand team as the first step in building a new design system.
AI features center stage.
Our novel AI features are now a first step in building a personalized fitness plan
HIGHLIGHT
Personalized
Fitness Plans
The plan is a commitment, but also supportive and flexible. It anticipates blockers and helps users build resilience as they work through them.
Key different-iator.
Our v.1 interactive workouts feature now core to the new experience
HIGHLIGHTS
Baseline interactive workouts
First workout calculates baseline fitness and form to inform workout progression
Checkin interactive workouts
Checkins track progress against the baselines. Makes adjustments as necessary on future workouts. Help keep motivation up.
12 weeks of work!
A celebration, final fitness report, and nudge for the next round
HIGHLIGHT
Personalized
Fitness Plans
The plan is a commitment, but also supportive and flexible. It anticipates blockers and helps users build resilience as they work through them.
What happened?
Going from evolution to revolution
v1
Increased CSAT, reduced churn
With personalization broken, designed a process to create ongoing, useful, fun, low-lift Feed content to address users' needs and frustrations
Increased app engagement
Interactive workouts became one of our most popular workout types.
v2
Aligned org on building a redesigned, refocused product
The AI Trainer concept and beta testing had very enthusiastic feedback.
Profitability time horizon proved too much for leadership
Andy Jassy became CEO and brought a more cold-eyed assessment to big bets like this one. He sunsetted the product.