Spotify
Timeline:
-
8 weeks (2023)​
Role:
Music streaming service platform
offers users to generate AI- curated playlists via voice or text input, customizing song selection based on mood, activity, or specific prompts - simplifying music discovery and enhancing engagement.
-
UX/UI Design
-
User Research
Tools:
-
Figma
-
FigJam
Problem
Users often feel overwhelmed when searching for new music due to an excess of recommendations and unclear discovery paths. The current system does not always align with individual listening preferences, resulting in a frustrating and inefficient browsing experience. As a result, engagement with the platform decreases, and users may seek alternative streaming services.
Project Goal
To improve the Spotify mobile experience by introducing an AI-driven playlist creation tool that allows users to:
-
Generate customized playlists using voice or text prompts tailored to their mood, activity, or specific music preferences.
-
Refine AI-generated suggestions by selecting favorite artists, adjusting playlist length, or removing unwanted songs.
-
Seamlessly integrate personalized playlists into their library for a more seamless and engaging listening experience.
​
Empathize
User Research
To gain insights into user behavior and pain points, I conducted user research through surveys, interviews, and usability testing. Key findings include:
-
User struggle to find relevant music for specific moods or activities​.
-
Many users are familiar with AI chatbots and express interest in a more interactive way to generate playlists.
-
Voice input for playlist creation is preferred for its convenience.
Desk Research
-
Personalization drives engagement, and AI-powered features provide a competitive advantage.​​​
-
Simplifying music discovery with intuitive features enhances user satisfaction.​
-
Transparency in data usage is essential for addressing privacy concerns.​
Project Goal
To enhance the Apartment.com mobile app by integrating a commute filter that enables users to :
-
Set commute preferences based on destination, mode of transport, and travel time.
-
View apartment listing filtered by commute times.
-
Compare multiple destinations (e.g., work and grocery store) within the app. This feature aims to improve the apartment search experience by reducing friction, improving decision-making, and increasing user engagement.
Define
Jobs-to-Be-Done (JTBD)
I created Jobs-to-Be-Done (JTBD) statements to focus on users' core needs and motivations, ensuring the AI-powered playlist generator directly addresses real pain points. By understanding how users discover, personalize, and engage with music, these statements guided feature development, ensuring a seamless, intuitive, and valuable experience within Spotify.
-
As a Spotify user, I want an AI-powered playlist generator to suggest new artists and genres based on my preferences, so I can easily discover fresh music without manually searching.
​
-
As a busy user, I want an AI assistant that generates playlists instantly based on my mood or activity, so I don't have to spend time searching for the right music.
​
-
As a hands-free user, I want to generate playlists using voice commands, so I can create music mixes effortlessly while multitasking.
Sketches
I used Crazy 8's to quickly explore UI ideas for the AI-powered playlist generator, ensuring a user-centered and engaging design for Spotify.
Design & Prototype
Hi-Fidelity Wireframes
Prototype
AI-Powered Playlist Assistant
Designed for effortless music discovery, the AI assistant allows users to create playlists instantly using voice or text prompts, customized to their mood, activity, or specific requests.
The content strategy prioritized user needs by refining input methods, optimizing AI responses, and ensuring a seamless, personalized listening experience.
Regenerating and Adding AI-Powered Playlists Results
This feature allows users to refine their AI-generated playlists by regenerating results with a single tap, ensuring that recommendations better match their mood, activity, or preferences.
​
By prioritizing clarity and control, the content strategy ensures an intuitive experience, allowing users to fine-tune playlists with minimal effort for a more personalized listening journey.
Test & Implement
Usability Testing
Number of Participants: 6
Interview Time: 30 - 50 minutes
Insights & Feedback
-
5 out of 6 users found the AI feature useful and time-saving
-
80% of users wanted a clearer way to regenerate specific results.
-
5 out of 6 users had difficulty finding the AI Assistant feature with the robot iconography in the Main Search Bar.
-
4 out of 6 users experienced that it would be nicer if they could also swap the results option when it is regenerated
-
90% of users were confused and not satisfied with the regenerating button on the keyboard search bar.
“I can easily see the button and there is no problem adding it. Very simple and easy step, love it!”
“It made sense that you’ll find the assistant in the search
region”
“I wasn’t sure if the AI received my input after what I said. It does not seem very smooth.”
“I am confused about the generate button. How will I know which input it is
regenerating when I click this button? What if I want to specifically regenerate a input?”
Iterations
Redesigning Regenerate Button
The regenerate button initially caused confusion, with 5 out of 6 users struggling to understand which input was being regenerated. Additionally, users found the button placement unclear, leading to difficulties in refining playlist results. To improve usability: ​
​
-
Redesigned the regenerate button to indicate which input is being modified.​
​
-
Relocated the button for better visibility and easier access.​
​​
Enhancing AI Processing Feedback
Users struggled to determine if the AI had received their input, with 70% experiencing uncertainty after submitting a voice or text command. To improve this issue:
​
-
Added a loading animation to visually indicate active processing.​
​
-
Refined visual and audio cues for a smoother, more responsive interaction.​
​​​
Conclusion
Next Steps
Moving forward, several improvements can be explored to refine the AI-powered playlist experience:​
​
-
Refining AI learning algorithms to deliver even more accurate and diverse recommendations over time.
-
Introducing collaborative playlist generation where multiple users can contribute prompts for shared listening experiences.
-
Integrating real-time feedback so users can instantly rate and adjust AI-generated playlists for better personalization.
Project Takeaways
This project was a deep dive into how AI can enhance music discovery while keeping the user experience simple and engaging. A few key lessons stood out:
​​​
-
Voice input is a game-changer. Many users preferred speaking their requests, making hands-free playlist creation a valuable addition.
-
Keep it simple. The easier it is to use, the more likely people are to engage with it.
-
Personalization matters. Users don't just want recommendations; they want control over their playlists.
​
This project reinforced the importance of building AI-powered features that feel intuitive, useful, and genuinely enhance how people experience music.
