Context

Timeline

2 months; Launched

Company

KKCompany

My Role

UI/UX Design

Motion Design

User Research

Team

PM*1

Engineering*2

Design*1 (Yit)

Introduction

KKCompany's flagship product is a cross-platform music streaming service that provides a seamless experience across mobile, web, and desktop. Primarily serving users in Asia, it has over 2 million paying subscribers.

Project Overview

The new feature, "Surprise Me" (SM), is a fully motion-driven AI playlist designed to offer users an innovative way to explore music. SM visualizes the algorithmic music generation process, allowing users to synchronize their auditory and visual senses to feel the rhythm.

As the lead for end-to-end product development, I managed both high-level strategy and hands-on design execution to launch the SM feature. My responsibilities included conducting user research, defining feature positioning through cross-functional workshops, and designing the user interface.

Problem

Users have lost confidence in our AI music playlists

Our algorithm received little attention in the past, leading to inconsistent recommendations. Users rely on their own playlists or search, contradicting our positioning as a music streaming service, where AI-driven discovery should be a core experience.

Solution

Rebranding the AI music recommendation experience

A new music discovery experience that:

  • Redefines users' perception of AI-generated playlists, ensuring past negative experiences no longer discourage engagement.
  • Aligns music discovery with our brand identity, enhancing user connection and acceptance of minor imperfections.
Jump to Final Design

Discovery

#1 "It takes too much time to find a suitable playlist to listen to!"

First, we conducted 6 user interviews to understand how users explore music and why they choose certain discovery methods.

Interview insights

#2 Provide graphic playlist covers for a more visually balanced layout

Compared to other music streaming services, our playlist covers rely heavily on album and artist images, while competitors use designed covers for better visual balance.

KKBOX vs. competitors: homepage comparison

The main insight

AI recommendations are judged as a whole, not by individual modules

Based on the trends in the demographics from the interviewees and my analysis:

  • The AI playlists on the homepage lack a cohesive design, reinforcing the perception of inaccuracy. Organizing them by complementary function can streamline exploration.
  • The typical playlist interface creates high cognitive load. We should diversify playlist formats to boost engagement.
  • Using user-focused playlist names makes recommendations feel more personalized.
Opportunity
How might we rebrand the AI music recommendation experience to drive higher user engagement?

Design

Co-creating workshops - brainstorming new ways to explore music

To explore ways to transform music discovery, we conducted a cross-functional workshop, bringing together product, engineering, and marketing teams to brainstorm ideas.

Workshop process

Creating a personalized music playlist gift

From our workshop, we aimed to repackage the AI recommendation experience in a simple, engaging way by designing a new playlist interface. Placing it at the bottom of the homepage, we saw it as a "special gift"—a fresh experience to re-engage users if earlier modules didn’t meet their expectations.

Strategy for feature placement

Iterations

Based on feedback, the designs continue to improve.

Old vs. new motion playlist

Delivery

"Surprise Me" interactive music playlist

"Listen, and be heard" is our brand slogan, which I used as the core concept. I translated this slogan into a motion graphic, showcasing how we listen to users' preferences and recommend the most suitable playlists.

Final design (Light/Dark Mode)

1. Engaging users through micro-motions

I added motion to the default background to capture users' attention and encourage immediate listening. It was also applied to error states to enhance user patience.

Micro-motions in specific states
Playlist cover

2. More Intuitive Recommendations

Use artist photos as covers to create a clearer connection. Users are more likely to recognize an artist's face, making it easier to decide whether they want to keep listening.

3. Visualizing the generation process

Transform the AI playlist generation into an immersive, brand-styled experience, allowing users to feel personally involved in the creation process.

Motion-based visualization

More details...

This new feature is designed for both mobile and larger screens.

Animations for Mobile and Larger Screens

Reflection

The perception of seconds can vary greatly among different users

After testing with team members, it was found that there was a significant variation in the acceptance of transition animation durations. After continuous testing, it was determined that most people found a completion time of within 4 seconds for the entire motion experience to be acceptable.

Convincing stakeholders and team members to try new experiments may require extra preparation

One of the biggest challenges in this project was determining the brand voice beyond just the product color. As a result, there had been numerous doubts and questions raised during the progress of the project. To address this, we had created countless prototypes and interviewed team members, asking them to imagine what the motion brand voice of our product would look like. This process had led to the final decision and resolution for this project.

Thanks for reading.