Accessibility and AI
Inclusive E-Commerce for Visually Impaired Users
Most shops are retrofitted for accessibility, if at all. This one started from the question of how someone who cannot see the screen would shop, and designed outward from there.
How the data moves
- Shopper inputScreen reader, voice, touch
- Accessible storefrontARIA and voice navigation
- Python backendFlask application
- AI product descriptionsTensorFlow
- Haptic feedbackConfirms actions by touch
Accessibility firstBuilt accessible from the first commit, not patched in afterwards.
What this covers
- A storefront built around screen reader support, using semantic markup and ARIA.
- Voice navigation, so a shopper can move through the shop without a pointer or screen.
- Haptic feedback as a non-visual signal during shopping.
- AI-powered product descriptions, so a product can be understood without seeing its photo.
Design decisions
Design for no screen first
Navigation, product detail and checkout were designed to work without seeing anything. Visual polish came after that worked.
Describe products with AI
Product photos carry most of the information in a shop. AI-powered descriptions give shoppers who cannot see them the same information.
More than one way in
Screen reader support, voice navigation and haptic feedback give shoppers several ways to use the site, so no single channel is a point of failure.