We were building an AI assistant that could not only give smart advice, but also help users transform their rooms — suggest decorations, rearrange layouts, and follow current interior design trends. My role as a product designer was to:
Challenge
- Design a chat interface that works smoothly on iOS, Android, macOS, Windows, and Web
- Handle differences between platforms (fonts, sizes, layout)
- And keep the user experience consistent and clear on all devices
This is me during the testing process.
Before jumping into design, I ran a quick research phase. I studied how users interact with AI tools in different platforms, looked into competitors, and defined the key expectations from an assistant in our product.
Process
Component System for All Platforms
- I created a universal chat component that worked across platforms:
- One structure, with platform-specific tweaks like font sizes and paddings.
- Fast switching between variants in Figma using well-organized components.
- Easy to maintain and scale — changes in logic or style update everywhere.
UX Challenges and Solutions
While working on the design, I faced a few challenges:
- On mobile, the chat blocked the editor — so I added auto-collapse logic based on user flow.
- I added haptics and small animations to improve feedback and interaction.
- I optimized behavior for slow networks — important for Web users in less stable regions.
- I designed smart scrolling and focus logic so users always stayed in context during long chats.
Optimizing the design files not only made the team work faster, but also improved the user experience. These changes helped increase subscriptions and reduce user churn.
Results and impact
- +19% subscription growth – users engaged more with the assistant and spent more time in the app.
- Return visits increased, especially from new users who found the AI assistant helpful.
- +36% interaction rate with AI – thanks to better accessibility across platforms.
- Fewer support requests – the assistant handled about 40% of common user questions
AI chat in website
