What if users were empowered to fine-tune Spotify's search algorithm to find music they love?

Tune is a chat feature that recognizes music discovery as an intentional, joyful activity worth designing for. Intended for people who love finding new music, but who're currently relying on passive recommendations and playlists from places like Instagram, YouTube, TikTok, and Discogs.

Tools: Figma, Lovable (Link to prototype)

Considerations
Realistic design - Ensure prototype looks realistic and consistent with Spotify's overall design system and tone of voice.

Time to value - Help users realize the feature's value quickly by creating an interactive, chatbot-led onboarding experience that demonstrates key use cases and the chat interface.

Onboarding
I initially designed an onboarding expereince that explained features using cards. Although beautiful and well explained, with short, snappy copy, I decided it was better to create a chat-based, interactive onboarding experience to familiarize users with the interface while ensuring the shortest possible time to value. (try it out here)

Learnings
Spend more time upfront perfecting design files to feed Lovable. - Correcting design output is tedious and eats tokens. Better to invest time on creating a designsystem early.

Design and perfect features/states/use cases separately before assembling. - Working on multiple at once appears to confuse Lovable and creates more bugs.

Name features and functionality early to avoid confusion. - Chatting with Lovable becomes easier if names and descriptions are consistent and laid out early. When features and workflows get complex, imprecise wording and naming lead to more bugs and low quality output.