AI Cover

Achieve a heavenly voice with just a simple recording and the power of AI.

PROBLEMS

External factors:

Online singing apps are becoming increasingly homogenized, and competition is intensifying.
The emergence of AI technology has raised the overall technical requirements. Only companies that keep up with the times can survive and develop.

Internal factors:

User stickiness is decreasing day by day. New functions need to be continuously introduced to attract and retain users.
The company's revenue has decreased. New business growth points need to be sought to increase profitability.

GOALS

Improve the recording quality at the smallest cost that users can recognize to enhance users' participation and willingness to share.

MY ROLE

This project is expected to be launched in two weeks. I am mainly responsible for designing the user interface of the product to ensure a smooth and intuitive user experience.
I and a PM jointly define the product's functions through methods such as user interviews and brainstorming. We also obtained the data provided by the data department for nearly three months to help us understand the real needs of the business and users.
I created a prototype to share the experience process, design principles, and implementation methods, which is conducive to reaching an agreement and promoting decision-making.
I explored a visual style suitable for the project, presented it, and output design specifications.
In the final testing stage, I tested and adjusted the displays of different models and platforms to ensure the unity of experience and vision.

SOLUTIONS

“Which function should be chosen for improvement as the breakthrough point?”
This is my first design challenge.

I remembered that when redesigning StarMaker App, through user research and data feedback, we learned that users are most concerned and dissatisfied with the recording quality of songs.
Recording, as the core function of StarMaker, is connected to multiple functions such as recommendation and community.
Therefore, by improving the recording experience, the overall improvement of upstream and downstream can be driven.
In recent years, with the development of AI technology, a relatively large improvement can be obtained at a relatively low cost.

Brainstorm：As a project starting from scratch, We selected two people from each participating department and four users to have a video meeting together.
Our goal is to further deeply understand users' visions for this product from different perspectives.

Insights：

Demand for top-notch recording quality!Users often skip sharing to the community if unhappy with recording. New users may even quit and cancel accounts. They want to be satisfied within twice.

Complicated custom sound effect modification tool.Users find it perplexing to manually modify the recording effect. They hope they can generate it with one click and share it directly.

Know What Users Need Better Than They Do.Users hope that AI Cover can understand them and provide personalized recommendatio

Discover：More than 60% of Starmaker's user group are middle-aged and elderly users. We originally thought that the use of new gameplay would only be highly attractive to young people. However, after research, to our surprise, when middle-aged and elderly users learned that the operation is simple and can improve the recording quality, they showed a stronger willingness to share.

Prototype v1.0

In Prototype v1.0, we had a setting where users could choose various timbres after recording. But in usability tests, users hesitated at this step, their emotions dropped, and some even quit without choosing. We realized it was choice overload. After discussion, we decided to cancel this function and only enhance recording quality through technical means. This is closer to our goal though it reduces some fun.

“I don't like traditional communication methods such as hand-drawn prototypes and sticky kanbans, which are extremely time-consuming in the process of agile development. ”

Throughout this project, I only choose paper prototypes when thinking alone, which is convenient for me to quickly record ideas. When ready to discuss with the team, I only provide online versions, which greatly improves efficiency and effectively retains records.

“It’s time to let it show!”

First abolished version.

After the usability test, the following feedback was collected.

Visual focus on top is confusing.
Main button lacks AI tech feel.
Vinyl record takes up too much space, reducing efficiency.

Final version

IMPACT

“Be pessimistic in mind but optimistic in action.”

Although this is an attempt in a new direction, the result has exceeded our expectations.
As of December 31, 2023, we were provided with：

Average user engagement duration: 2’30”;
Feature success rate: 92%;
Average processing time: 3.2 seconds;
Subscription conversion rate driven by the feature: 8%.
New feature iteration increases user engagement by: 15% almost

REFLECTIONS

“Understanding users should be a habit.”

Although the data results seem good, looking back at the beginning, we were still hesitant about whether to start with the recording function. It seems that we still don't know our users well enough.

“This Is a Try of Achieving Big with Little.”

This is an evaluation given by our group's CEO.

“No dumb questions”

During the project, design biases persisted. Initially, we wrongly assumed that new technologies like AI only attract young people. Thankfully, research results corrected this notion and guided us to balance song resources across age groups in the later stage.

Today, there are already tens of millions of users using AI Cover. Successful designs require time to polish. For the design of AI products, clearer prompts, simpler operations, and faster output are our goals to pursue in the future.

Relevant videos
🕒 Introduction at 2:51

Thanks for watching