Pet project · UXcel Top 6 Global

Spotify Symphony

Role
Sole designer (pet project)
Timeline
2024 concept sprint
Stack
Figma prototype
Result
#6 globally, UXcel Design Challenge
HERO IMAGE
Cover — the Symphony Chat interface with a Taste Map visualization
01
The challenge

Design Spotify's AI music companion — without algorithmic overload.

This was a pet project built for the UXcel Design Challenge: design an AI music companion for Spotify that helps people discover new music without just handing them another passive algorithmic feed. Discover Weekly and Daily Mixes are good — but they're one-directional. There's no back-and-forth, and no sense of why a recommendation showed up.

My angle: music discovery isn't consumption, it's conversation. People don't just want to be told what to like — they want to explore alongside something that's actually listening.

Category
Conceptual · Personal project
Challenge
UXcel Design Challenge
Result
Ranked #6 globally
Format
Figma prototype + case study
02
The concept

Symphony — a companion that treats discovery as dialogue, not delivery.

Feature
Taste Maps
An interactive visualization of a listener's taste as clusters — lo-fi beats, 90s hip-hop, indie folk — that they can drag to say 'more like this, less like that,' updating recommendations live.
Feature
Why This?
Every recommendation carries a short explainer — e.g. 'you saved three Frank Ocean songs; this producer worked with him' — turning the feed into something that teaches a listener about their own taste.
Feature
Symphony Chat
Natural-language requests instead of browsing — 'find me sad songs for a rainy Tuesday,' or 'music for a long coding session' — as a second way in alongside the visual map.
Feature
Collaborative Playlists
Human-and-AI co-curation on a shared playlist, where the AI suggests and the human decides, with a visible trail of who added what.
ARTIFACT
Taste Map visualization next to the Symphony Chat conversational interface
03
Design decisions

A companion with a voice, not just a feature set.

Symphony's personality was deliberately friendly and a little sarcastic — a music-nerd tone rather than a neutral assistant voice. Visuals leaned on music waves and album art instead of generic AI-product visual language, and onboarding asked people to "tell me your music story" instead of building a dry preference profile.

"Music discovery isn't about consumption. It's about conversation."
The core reframe
04
What I'd improve

Honest gaps, looking back.

As a pet project, the concept leaned harder into the creative angle than into rigor I'd bring to client work — I'd want to think harder about accessibility and edge cases (what does Symphony do for someone actively skeptical of AI recommendations?), and I'd want real usage data rather than a prototype to know whether "Why This?" actually changes behavior or just reads well in a case study.

05
Outcome

Ranked #6 globally in the UXcel Design Challenge.

The concept placed #6 out of the challenge's global cohort — the one hard number from this project I can point to directly. Everything else here is a conceptual exploration, not a shipped, measured product; I'd rather be clear about that than dress it up as more than it is.

We don't just want to be told what to like.
We want to explore together.

Create a free website with Framer, the website builder loved by startups, designers and agencies.