Helping weightlifters safely train solo by using AI-driven tech to provide real-time form corrections.
At a glance
Spot Me is a weightlifting training app using AI-driven pose estimation to "spot" your form. It gives real-time feedback, so you can focus on lifting your best without worrying about getting hurt. As the sole product designer for my capstone project, I was inspired by hearing about experiences of weightlifters seriously injuring themselves because they didn't know how to execute a move correctly while working out alone. By creating a digital tool, I aimed to design a digital solution that was accessible and improved the solo lifting experience.
Role
Principle Designer
Timeline
4 months
Area
Mobile App
Context
Incorrect form during weightlifting can lead to serious injury.
There is no significant association between injury occurrence and gender, age, or body mass index. However, there is a significant association between the occurrence of injury and weight carried while lifting weights. Since casual lifters struggle to correct their form when lifting alone but continue increasing the weight carried, they risk serious injury. (NHI, 2023)
Why is this problem important?
1 in 4
People reported doing weight-lifting activities weekly.
CDC, 2022
61%
Of weightlifting injuries are due to technical errors.
CBI Insight
Research & User Interviews
What is the current weightlifting experience?
To create a solution that would really help people, I first wanted to understand the current lifting experience. I conducted 7 preliminary interviews with a wide range of users, from beginners (< 1 year) to veterans (10+ years) and those who don't use any tech to power users. Based on my insights, I identified 3 key pain points which informed my ideation process.
Who am I designing for?
The Lifting Newbie
Weightlifting for 3 months
Behaviors
Just began lifting and following fitness influencer workout videos. They worry about getting hurt when trying a new set but find the personal trainers too expensive to start off with.
Goals
To get personalized guidance during workouts to make sure their form is correct and prevent injury.
The Dedicated Learner
Weightlifting for 5+ years
Behaviors
Working out is an essential part of their routine. They started researching proper form after a serious injury, not realizing they'd been doing the exercise incorrectly the whole time.
Goals
To consistently improve their form to optimize progress and see results like a new 'personal best' in weight carried.
Turning insights into design challenges
Problem Space #1
It's difficult to learn the correct form by yourself.
"I was training my chest and kept increasing the weight but didn't know the right degree to do it in. I ended up getting a serious injury."
Dennis, Weightlifting 10 Years
How might we teach lifters the correct form?
Problem Space #2
Lifters desire personal guidance, but find trainers too expensive.
"I want a personal trainer to teach me, but I can't afford one right now."
Hao, Weightlifting 2 years
How might we digitally provide correctional form feedback?
Problem Space #3
Doing exercises by memory can still lead to incorrect form.
"I was doing bench presses wrong for a year. Just by memory, you can still have incorrect form."
priyana, Weightlifting 7 Years
How might we encourage continuous form improvement?
Divergent Concepts
How will I address the pain points in my design?
I developed 4 divergent concepts to test with users and elicit strong reactions from them. I created low-fidelity prototypes to focus their reactions on the ideas rather than the design so that I could identify which solutions resonated with them and which to leave behind.
Concept 1:
Biometric Tracking for Injury Prevention
When abnormal biometrics are detected by the user's smartwatch, the routine pauses.
Concept 2:
360° Viewable Poses
Users can see all angles of the proper form instead of just a video from 1 point of view.
Concept 3:
Responsive, Visual Feedback
Smart mirror tracks and analyses the user's form, providing real-time corrections.
What do users think?
Users were most drawn to real-time coaching and injury prevention.
These were the things they struggled with the most when working out alone. They wouldn't invest in a smart mirror but wanted the form correction it provided. I aimed to design a tool that could act as the "next best thing" to having a personal trainer, since so many lifters I talked to felt trainers were essential but often too expensive. By creating a digital product, I could make these services more accessible and affordable, so more people could benefit.
Narrowing the Scope
'Spotting' in weight training is supporting the person lifting.
Spotters help the athlete complete the lift, ensure proper technique, and prevent injury. They also boost an athlete’s confidence level and morale, leading to better performance during their workout.
Could we 'spot' weightlifters digitally?
Human pose estimation for corrective feedback.
HPE gets the most attention in AI fitness applications since it can analyze the movement of athletes in different scenarios using just a smartphone camera.
Identifying opportunities in the user journey
Outlining key features
Tailoring the product to the weightlifting experience
Track Users Form During Workouts
Do users want to choose when they're being tracked, or would that be the default?
Alerts Ensure Safe, Optimal Lifting
How much information on form analysis would be the most helpful for users and when?
Personalized Progress Tracking for Motivation
Will users want to compare the day's workout results with their progress history?
Ideation
Developing a product that teaches weightlifters the proper form
I started by sketching out different screens based on insights from my interviewees, including coach-led workout routines, form analysis and corrections, and post-workout analysis. To gain perspective on which screens were missing, important, or unnecessary, I created a system map outlining the correction feedback and post-workout flow.
Sketching
Home Page, Weekly Progress
Workout Video
Human Pose Estimation Tracking
Post-Workout Analysis
Design Development
Refining the UI through key design iterations and decisions
I turned these sketches into mid-fidelity prototypes to enhance my UI elements. I tested them with casual weightlifters who had just started lifting and those who had been doing it for years to ensure the UX could easily be introduced into their routine.
Mid-Fidelity Prototype
Design Consideration #1
Improved navigability from Coach to Spot Me mode
Main Issues
Lifters prioritize quality over speed, so time indicating UI is unnecessary.
Rather than taking videos, users prefer live analysis.
Design Solutions
Remove the timer and progress bar.
Highlight exercise name, reps, and weights.
Toggle easily switches between modes for feedback.
Design Consideration #2
Enhancing the readability and legibility of alerts
Main Issues
Need real-time feedback for immediate correction.
Alert size and text were too small to read from afar.
Prefers to read further information after the workout.
Design Solutions
Form alerts are large, simple, and audible.
Fullscreen modal to capture the user's attention.
An in-depth form analysis is given post-workout.
Design Consideration #3
Focusing on personalized progress tracking
Main Issues
Stats are too general- users want a set-by-set accuracy breakdown.
Helpful to see the progress over different periods.
Design Solutions
Show progress over week, month, and year.
Highlight improvements in accuracy, alerts, and weights lifted.
Final Design
A weightlifting training app that leverages AI to analyze your form and provide real-time correctional feedback.
Step-by-step Workout Guide
How might we teach lifters the correct form?
Easily track this week's progress, motivating you to keep up the hustle. Workout cards and coach-led video tutorials make it easy to follow along each set and alleviate the stress of routine planning.
Pose Analysis for Form Feedback
How might we provide correctional feedback on their form?
Easily toggle between coach-led video tutorials and 'Spot Me' pose tracking mode. When incorrect form is detected, it triggers bite-size audio and visual alerts on how to correct yourself- as a coach would! Modal disappears once your form has been safely corrected.
Tracking Instructions
Human Pose Estimation
Personalized Progress Tracking
How might we encourage lifters to keep improving their form?
End the workout with a brief recap on how you did and gives an option for detailed information on each set's form accuracy. Lets you see all alerts to understand your mistakes and tips on how to improve.
Beyond today's workout, you can track your progress with that set over time including accuracy, alerts, and weight carried. Highlights even the smallest statistical improvements since it often takes longer to see changes physically.
Alerts Recap
Progress Over Time
Reflections
What did I learn?
Design Isn't a Linear Process
At the start of this project, I was uncertain about which fitness problem to tackle. It wasn’t until my mid-fidelity interviews that I focused on weightlifting. The process involved revisiting my research, but that’s the thrill of design— drawing, researching, testing, and repeating. Keeping the user’s needs as my north star guided me to a genuine area of opportunity.
Don't Forget About The Tech!
I love getting into UI and UX design, but that’s just one piece of the puzzle. Researching AI has shown me how crucial it is to stay updated with the latest tech to keep innovating. I think it’s all about balancing the tech we have with imagining what it could be to keep pushing the boundaries.
Kirsten Geiger
Work
About
Resume
Let's get in touch!
Work
Spot Me
Vocal Mind
Summit Learning
Say Hello
kirstengeigerdesign@gmail.com
© Kirsten Geiger 2024