Microsoft Seeing AI Redesign

OVERVIEW
Microsoft Seeing AI is an intelligent camera app designed for the blind and low vision community that provides information about who and what is around you. Current users who have experienced and benefited from the app also face many challenges and frustrations when using it.

DURATION
10 weeks

PROJECT TYPE
Personal

TOOLS
Lucidchart, Sketch

PROBLEM STATEMENT
How might we improve the user experience and accessibility of Microsoft Seeing AI?
DESIGN PROCESS
WHO ARE THE USERS?​​​​​​​
Our users struggle with low vision and seek to use technology to ease their difficulties daily and making them more independent. 
Meet Haley
45, college admissions counselor
Haley came from a well-educated middle-class family. She was diagnosed with glaucoma then started experiencing a decline in vision, and it had become worse ever since. She now lives a much simpler life and likes to borrow a pair of eyes from her husband, Ben. ​​​​​​​Haley has her prescription glasses with her all the time. One of her best friends is the magnifier feature on her phone and iPad. Haley seeks assistive technology to help her improve the quality of daily life.
"Just because something can't be done to correct our vision doesn't mean there isn't something that will enhance our lives."
COMPETITOR ANALYSIS
The most widely used product in the market for people with low vision is the traditional assistive device, magnifiers that come in different forms and sizes.
After putting together the competitor analysis, I was able to find the insight: shortening the learning curve of the Seeing AI app and making it more adaptable.
RESEARCH
To understand the experience from our user's perspective, I conducted secondary research and primary research using surveys, interviews, fly-on-the-wall observations. I had participants walk through the experience of using the current app, asking for their options on what works and what doesn't work. 
PAINPOINTS
I received consistent feedback from a majority of participates. The feedbacks include immediate needs and next updates based on feature prioritization. The two frustrations below should be priorities to improve.
1. The placement for scanning is challenging and always shows error
"The struggle is to capture text and not knowing why it can't recognize."
"Document scanning depends on the correct placement, but if I can't see clearly, how do I accurately place these ?"
2. Hard to navigate with buttons close to each other
"Pause button is too close to the channel bar. The accidental press is too easy." 
SOLUTIONS
The frustrations current Seeing AI users face when using the product can be resolved by upgrading features including: 
Descriptive error messages
such as "too dark," "too close," "move more to the right" to help users who have low vision better scan the object.
Accessible gestures
such as shaking the phone to cancel scanning, swipe left and right with two fingers to navigate different features, etc.
SCENARIO
FLOWCHART
UI DECISIONS
Along with the hypothesis, I took this opportunity to revisit the UI within the app. I was explicitly focusing on implementing the best practices for inclusive design. 
CURRENT APP ANALYSIS
BEST PRACTICES FOR INCLUSIVE DESIGN
I followed Apple's Human Interface Guideline in accessibility when making the UI updates.
• Reduced transparency
• Guide the user with VoiceOver
• Increased text size
• Use regular or heavy font weights
• Avoid using italics and all caps
ONBOARDING EXPERIENCE
Onboarding videos will automatically play when the user navigates to each feature for the first time to help shorten the app's learning curve. It'll also be available on the right-hand corner afterward for the user to refresh their memories at any time.
B/W WIREFRAMES
Onboarding Video
Onboarding Video
Person Recognition
Person Recognition
Remember This Person?
Remember This Person?
Enter a Name
Enter a Name
Sign Up / Sign In
Sign Up / Sign In
Success!
Success!
USER TESTING & FEEDBACK
I tested the redesigned experience with users to see if there were any more opportunities for improvement. I brought the same participants from the initial user interviews to test the redesigned version. ​​​​​​​
SAMPLE USER FLOW
USER FEEDBACK
The response I got from users was that it was much easier and more effective to use the app in the redesigned experience. The redesign helped establish confidence among users to use Seeing AI. 
FUTURE CONSIDERATIONS
During the user research, I noticed repetitive pain points from users that were not immediate needs but were worth improving as the next steps for the app redesign.  
1. Lighting detection
There is an existing light feature that generates an audible tone corresponding to the brightness in your surroundings. Most users say this feature has more potential to be helpful than what it does now. 
With this upgrade, users will find lighting sources in the space to use as a landmark when learning the new space's layout through audible tone guidance.
​​​​​​​
2. 3D space mapping
The new upgrade of the scene feature will help the user detect relevant information in the space such as empty seats, trash can, AND restroom as users move the phone around the space. ​​​​​​​
TAKEAWAY
The power of co-designing with users
When conducting user interviews, I discovered users shared their opinions on the app experiences and brought up many exciting ideas on how to improve features or how they wished the app could include certain new features. It occurred to me that users were unintentionally co-designing with me. As experts of their own experiences, co-designing with users can bring different views that inform design and innovation direction whose ultimate goal is to serve the same users group best. For future projects, I wish to formally bring in users to co-design with me during the ideation and prototyping process and explore different co-design methods.
Back to Top