AI covers for playlists

This case study showcases how playlists were reimagined through an innovative AI image generating platform and the application of behavioural design principles to the user experience.

I used AI to help me write up this case study. Can you tell where? 😉

To play, press and hold the enter key. To stop, release the enter key.

Project overview

Summary & team

Our team worked with a popular music streaming service to create a microsite that allows Gen Z users to express themselves through their playlists with AI image generation.

This project was a collaborative team effort with myself as UX designer, a project manager, a visual designer, a motion graphics designer, a behavioural strategist and a tech lead.

Goal & hypothesis

To engage Gen Z users, we designed an AI-powered feature that creates unique imagery from playlist metadata. This innovation aimed to boost social sharing and expand the platform's reach within the target demographic.

Our hypothesis was that if we make the experience simple while grounding our decisions on behavioural science principles, users will find it intuitive and be nudged to complete the whole process.

Process

The process included defining behavioural design principles, creating seamless user flows, wireframing, testing the designs with Gen Z users who gave invaluable feedback, and finally adding some style with UI & motion design.

Final website walkthrough

Principles

Laying a behavioural science foundation

Here are the main 3 behavioural science principles we implemented to support our hypothesis:

The Endowment Effect

Users are more likely to want to keep something that they own.

We activated users’ sense of ownership by labelling themselves as the artist, i.e. [Playlist name]’s cover art by [Username]

Visible Effort

Users are more likely to want to keep something that they own.
Revealing to users what is going on ‘behind the scenes’ of a process has been found to increase their valuation of products.

We included a series of quirky loading screens to let the user know what the AI is up to.

Variable Reward

Users have a deep, hidden love of unpredictable rewards. They feel overly excited to receive surprises, and love the anticipation of ‘games of chance’.

We added a “Rolling the dice” effect where regenerating means the user will lose their current art.

What we learned from testing with Gen Z

We tested with 5 participants between the ages of 21-27 through moderated usability testing of a mid-fidelity prototype.

Playlist covers are personal

All participants personalise their playlists through imagery that reflects the mood of the playlist or their current mental life phase. Through this we learned that the chosen art direction of the generated imagery might not match the personalities of all Gen Z users.

AI is highly engaging

Participants were excited about the novelty of the experience and they're looking forward to add their personalities to their playlists.

Fun loading screens decrease perceived waiting time

Users were truly delighted by the humour that was worked into the loading screens. On average, they thought they were waiting around 3-5 seconds, when it took around 10 to generate their cart. We also learned that they expect a variety of screens to avoid monotony which meant a logic had to be built into the system that varies the sequence of the loading screens.

Copy is king

There was some vagueness around how the generation works and users wanted to be more in the know about what data of theirs are impacting the art. This lead us to create a 'how it works' section.