Design Exploration
Testing Stitch AI.
Google recently introduced Stitch AI, a tool that generates interfaces from prompts. Instead of recreating an existing app, I used it to explore an idea I'd been thinking about for quite some time — a product for India's motorcycle riding community.
The Idea
I wanted to see how well Stitch could understand a product idea rather than a UI prompt.
As someone who enjoys motorcycle touring, I've always felt riders rely on too many disconnected platforms. Routes come from Maps, ride planning happens in WhatsApp, road updates live inside old Reddit posts, and discovering new rides depends on social media.
Instead of designing another maps app, I imagined a community-first product where riders could discover routes, organize rides and stay informed in one place.
I deliberately gave Stitch the freedom to decide the visual language, layout and design system.
What Stitch gave me
What surprised me
Consistency
Every screen followed the same visual language without me defining one.
Design System
It generated colours, typography, spacing and reusable UI patterns that felt surprisingly coherent.
Structure
Instead of random screens, Stitch understood navigation and produced interfaces that felt connected as part of one product.
Speed
Going from an idea to something visual took minutes instead of hours, making it much easier to evaluate whether the concept had potential.
Final Thoughts
Stitch surprised me more than I expected.
It didn't just generate individual screens — it generated a surprisingly coherent visual language with colors, typography and reusable components.
Where it still struggles is product thinking. Interactions involving maps, rider collaboration and more complex workflows still need human decisions. Some screens also felt generic and would need refinement before becoming a real product.
For me, tools like Stitch don't replace design. They simply make the first step much faster.