SONO: AI Music Discovery Site

Reimagining music discovery through emotion, conversation, and personalization

Designing a Mood-Based Music Discovery Experience

At the start of our capstone year, our group worked on a brand redesign for a small business called Blush & Co. This taught us how to communicate with a real client, organize research, and build a visual identity. When the fall semester began, we decided to take what we learned and pivot toward an idea that felt more personally exciting and creatively challenging. That idea became SONO, a mood-driven music discovery platform built around emotional expression rather than traditional recommendation algorithms.

SONO gave our team the chance to explore a fully original product concept and imagine a new way for people to find music that reflects their feelings and energy in real time.

Homescreen of SONO

Project Date

Fall 2025 — Spring 2026

About SONO

SONO is a mood-based music discovery platform designed to help users find music that matches how they feel in the moment. Instead of relying only on listening history or algorithm-driven playlists, it takes a more human-centered approach by letting users express their emotions and receive music that reflects those feelings in real time.

What started as a simple mood-based concept evolved into a more interactive, conversational experience powered by AI. SONO guides users through intuitive emotional prompts, color cues, and mood selections, making the process feel natural and expressive rather than data-driven. The goal is to shift music discovery away from passive recommendations and toward something that feels more intentional and personal.

The platform is designed for anyone who wants music that aligns with their current state, whether they’re trying to focus, unwind, recharge, or feel inspired. Overall, SONO reimagines music discovery by centering it around mood and vibe, creating an experience that feels both personal and easy to use.

Our Roles

Throughout the SONO project, we worked as a fully collaborative team. Even though each of us had main areas we led, we all contributed to coding, design, research, user testing, and presentations. Our roles stayed flexible as the project evolved, and we made sure the final product felt like something we truly built together.

I led overall development, focusing on frontend engineering and UX design. I built the core interfaces, set up the GitHub repository, and ensured everything stayed accessible and visually consistent. I planned features and set weekly development deadlines. I guided our user flows, collaborated closely with Jamar on feature integration, and helped with testing and debugging throughout the semester.

Jamar led backend development and graphic design. He took charge of building functionality behind the scenes and worked closely with me to ensure everything connected smoothly and realistically from a technical standpoint. He also helped implement the AI chatbot feature and integrated the Spotify API into our app as the concept grew. He also designed all logos and asset materials.

Sa’Nya served as our Project Manager and originally came up with the idea for SONO. She kept us organized, helped set deadlines, guided our weekly check-ins, and made sure our deliverables stayed on track. She handled communication with instructors on Slack, led much of the structure for our presentations on Canva and Capcut, and supported both the creative sides of the project.

Evie supported UX, research, and branding. She handled much of the presentation and styling work, including most of our Canva decks and the UX Strategy Toolkit. She also added the AI chatbot concept that expanded the original idea and helped shape the direction of how SONO could feel more interactive. Along the way, Evie contributed to competitive research, early UI ideas, and the overall visual tone of the brand.

Together, these roles allowed us to blend our strengths, support one another, and build a project that felt collaborative from start to finish.

The Problem

Music discovery today is built around data. It focuses on listening history, trends, and patterns.

But emotions do not work that way.

People often know how they feel but do not know what to search for. You might feel overwhelmed, nostalgic, or somewhere in between, and there is no clear system that captures that nuance.

Our challenge was to design a product that:

• translates emotion into meaningful music recommendations
• feels intuitive and natural to use
• moves beyond generic, algorithm-based discovery

At the same time, we had to balance technical limitations, user expectations shaped by platforms like Spotify, and the complexity of building something that feels personal instead of automated.

Most music platforms rely on behavior-based recommendations rather than emotional input.

Competitive Research

To understand where SONO fits within the larger landscape of music discovery tools, our team analyzed both direct competitors and indirect competitors.

Direct Competitors

These platforms focus specifically on alternative music discovery methods, emotional selection, or algorithm-light exploration:

• Gnoosic
• Skiley
• Every Noise at Once
• Musicovery
• Music Map

These tools helped us understand what already exists in the emotional and exploratory discovery space. Many of them prioritize genre mapping or minimalistic suggestion models. SONO builds on these ideas by adding a more expressive, personality-driven interface and making the experience visually and emotionally engaging.

Indirect Competitors

These major platforms influence the broader streaming and discovery landscape, even though SONO is not trying to replicate them:

• Spotify
• Apple Music
• YouTube Music
• SoundCloud
• Tidal

We referenced these platforms for UI clarity, accessibility standards, and overall user expectations. While they are powerful at large-scale recommendation, their discovery tools remain centered on user history rather than emotion. SONO positions itself as a complementary experience rather than a replacement.

The Solution:

We developed SONO, a platform centered around emotion-first discovery. Instead of starting with genres or artists, users begin with how they feel. The goal was to shift music discovery from something users search for into something that responds to them.

Key Features

SONO’s prototype highlights a set of core features designed for simplicity and expression:

• Fast onboarding flow with Spotify integration
• Mood Calendar to track the user’s mood and playing history
• Badges/achievements to create incentives to encourage users to return to our app daily and track progress
• An AI chatbot that adjusts recommendations based on emotional inputs and conversation
• A clean, colorful interface for saving songs to the user’s Spotify account
• Social sharing tools that make it easy to send moods to friends

The Process

Our first step was defining the core identity of SONO, which began with building our UX Strategy Toolkit. This toolkit became the central place where we organized everything from visual moodboards and logo sketches to color exploration, UI inspirations, and early wireframes for the onboarding and recommendation flows. It also helped us outline our user personas, experience goals, and overall strategy for how SONO should feel and function.

From there, we created a preview deck that summarized the most important pieces of the toolkit, including:

• SONO’s purpose and value
• Mood and aesthetic direction
• Feature breakdown
• Technical needs
• Competitive landscape
• Research insights
• Our timeline and deliverables

The toolkit guided nearly all of our design and decision-making. Throughout the semester, we met weekly to update it, refine UI concepts, revise written content, and prepare the materials that eventually became our pitch deck and final demo. It acted as a living document that kept our project structured, focused, and aligned from start to finish.

Our Toolkit Introduction Page

Our Design System

The design of SONO was built to support this more emotional and conversational experience. Colors were tied directly to moods, helping users quickly connect visually with how they were feeling. The voice of the platform was intentionally conversational and human, avoiding anything that felt overly technical or robotic. Messaging focused on emotion rather than explanation, keeping the experience simple and relatable.

Everything worked together to make the platform feel intuitive and expressive, so users could focus on describing their feelings rather than figuring out how to use the product.

Our design system colors and messaging behind them

Our design system fonts and the messaging behind them

Early Prototype

Our first SONO prototype created on Figma

Originally, SONO was much simpler.

The idea focused on a mood-based interface where users could select how they felt and receive music that matched that mood. It was visual, intuitive, and easy to understand.

But as we started building it, something felt missing.

Mood selection worked at a surface level, but it did not capture the complexity of how people actually feel.

The Big Moment: Introduction of AI

The introduction of Aria shifted SONO from a selection-based tool to a conversational experience.

AI was not part of the original plan for SONO. At first, we saw it as a way to enhance personalization, but as we explored it more, it started to reshape the entire experience. Instead of selecting a mood from preset options, users could describe how they felt in their own words. This shifted the platform into something more conversational and expressive, where the interaction felt less like choosing and more like being understood.

This change was one of the most exciting parts of the project, but it also introduced a new level of complexity. Once AI became central, expectations changed. It was no longer enough for the app to simply work; it needed to respond in a way that felt thoughtful and emotionally aligned with the user. People weren’t just looking for recommendations; they were expecting the system to reflect what they said and how they felt.

That’s where we started to see gaps between the idea and the experience. While the concept of AI-driven, emotion-based discovery was compelling, consistently delivering responses that felt accurate and meaningful proved to be much more challenging.

User Research and Testing

After building out our 1.0 of SONO, we conducted two rounds of usability testing with multiple participants. At first, the results seemed positive. Users were able to move through the app easily, navigation felt clear, and tasks were completed without confusion. From a usability standpoint, everything was working as intended.

But as we continued testing, a different pattern started to come up. Users weren’t confused; they were underwhelmed. Even though they could complete tasks, their reactions pointed to something missing. Comments like “it feels kind of generic” or “this is stuff I already know” made it clear that the experience wasn’t leaving an impact.

This became a turning point for the project. We realized that usability alone wasn’t enough. SONO functioned well, but it didn’t feel meaningful yet.

The biggest gap showed up in the AI experience. Users expected a conversation, but what they got felt more like quick outputs. Aria would often jump straight into recommendations without asking follow-up questions or really building on what the user said. One piece of feedback summed it up well: “It’s a chatbot. Let it chat.” There was a clear mismatch between expectation and reality.

Because SONO is centered around emotion, this disconnect felt even more noticeable. When recommendations didn’t align with what users expressed, it didn’t just feel slightly off; it felt like the system wasn’t really understanding them.

At the same time, there were still strong positives. Users consistently liked the overall concept and the idea of expressing their feelings to discover music. Many said they would use it if it felt more personalized. The Mood Calendar, even without full functionality, was one of the most well-received features. It showed us that the foundation was there, but the experience needed to go further to fully deliver on that idea.

Iteration and Improvements

After testing, we made several key improvements focused on closing the gap between the idea and the actual experience.

We first worked on strengthening the AI interaction by making it feel more conversational. Instead of jumping straight into recommendations, we introduced more flow and better connections between what the user said and what was returned, so it felt more like a back-and-forth rather than a one-step output.

We also expanded music discovery by increasing the range of results and reducing repetition, since one of the biggest issues was recommendations feeling too predictable or familiar.

On top of that, we refined personalization to better align user mood input with the recommendations. The goal was to make responses feel more intentional and emotionally accurate, not just loosely related.

Finally, we revisited features like the Mood Calendar to improve their clarity and purpose, making sure users understood how and why they would use them.

Together, these changes helped move the experience closer to what users expected, making it feel more personal, dynamic, and aligned with the original vision.

Our Biggest Hurdle

One of the biggest struggles we encountered came from external limitations that were out of our control. In early March, Spotify updated its developer policies, which significantly impacted both our testing process and the level of personalization we could achieve. Access to the app was restricted to a small number of approved users (around 5 at a time), all of whom had to be manually added by email. On top of that, users needed a Spotify Premium account to fully access certain features. This made it difficult to run flexible usability testing, since participants couldn’t freely log in and experience the product on their own.

At the same time, Spotify also limited the type of user data we could access. We had less visibility into personalized listening behavior, which meant we couldn’t rely as much on individual preferences or history to drive recommendations. As a result, the outputs leaned more generic, which directly conflicted with our goal of creating something that felt emotionally personal and tailored.

Because these challenges happened in the middle of testing, we had to adapt quickly. We shifted part of our AI approach from the OpenAI API to Anthropic’s Claude model, aiming to make the experience more intuitive and conversational despite the data limitations. This wasn’t part of the original plan, but it allowed us to focus more on how users expressed their emotions and how the system responded, rather than relying heavily on backend data.

Overall, these constraints added a layer of complexity to the project. They limited both who could access the app and how personalized the experience could be, pushing us to rethink our approach and find creative ways to stay aligned with our original vision.

Results

By the end of the project, SONO evolved into a fully developed prototype that:

• demonstrated a new, emotion-first approach to music discovery
• highlighted how conversational AI can shape user experience expectations
• validated strong user interest in mood-based discovery, even with technical limitations

More importantly, the project revealed that usability alone isn’t enough. For products centered around emotion, perceived understanding and personalization are what define success.

Homepage: Where users land when opening the site

Homepage

Aria, AI Chatbot: Start with Aria, who helps you discover your emotions through natural conversation to help you discover music based on how you are feeling

“Aria,” AI Chatbot Recommendation, Page

Recommendations: Based on how you are feeling, Aria shows you music to add to your Spotify library. Have a quick preview of each sound or refresh if they are not to your liking. You can add songs to your Spotify library by clicking on the plus icon.

Recommendations from Aria

Mood Calendar: Keep track of your moods and achievements based on how often you use SONO. Gamifying our site incentivizes users to come back.

Mood Calendar Page

SONO tracks your daily mood history based on conversations with Aria

Go back and explore previous moods and recommendations whenever you like!

Keep track of your acheivements!

Trailer:

The SONO trailer was created to capture the emotional core of the product and communicate its value in a quick, engaging way. Rather than focusing on technical features, the trailer emphasizes the feeling of discovering music that truly matches your mood. It highlights the shift from traditional, data-driven recommendations to a more expressive, human-centered experience where users can describe how they feel and receive music in return.

Through a combination of visuals, pacing, and audio, the trailer mirrors the experience of using SONO itself. It guides the viewer from emotion to interaction to discovery, showing how the platform responds to users in a way that feels intuitive and personal. This also served as a way to practice storytelling for our final presentation, helping us translate our product into a narrative that resonates beyond just the interface.

Watch our trailer: https://youtu.be/w7TzrdPIrtQ?si=gp9RioKLJ3QuKKzj

Poster:

The SONO poster was designed as a visual extension of the product’s core idea, communicating emotion-driven music discovery in a bold and immediate way. The central message, “Find music that fits your mood,” anchors the design and clearly conveys the value of the platform at a glance. Rather than focusing on features, the poster emphasizes feeling, aligning with SONO’s goal of making music discovery more expressive and personal.

The design system plays a key role in reinforcing this. The gradient color palette reflects shifting moods, while the wave elements subtly represent sound and emotional flow. The use of a relaxed, immersed subject visually captures the experience of connecting with music on a deeper level. Typography is bold and simple, making the message easy to read while maintaining a strong visual hierarchy.

We also incorporated a QR code to call users to action, to bridge the physical and digital experience, allowing viewers to directly engage with the product. Overall, the poster balances branding, emotion, and clarity, serving as both a promotional piece and a concise representation of SONO’s identity.

Our 4k Poster

Our Square Poster

Reflection

SONO is more than just another music tool. It is a human-centered approach to music discovery that encourages people to choose sound based on how they feel. This project allowed us to combine design thinking, UX research, emotional branding, and storytelling into a single concept.

Pivoting from Blush & Co., SONO opened the door for more creative exploration and pushed us to grow as designers, collaborators, and strategists. Building SONO was both challenging and rewarding, and it became the most meaningful project of our semester.

Looking ahead, we hope to continue developing SONO beyond its current limitations. Our goal is to scale the platform so it can access more personalized data, deliver more intuitive interactions, and generate results that feel truly aligned with each user. We want the AI to feel more conversational and emotionally aware, with stronger follow-up and deeper understanding of user input.

Future goals include improving personalization, expanding music discovery to feel less repetitive, and fully developing features like the Mood Calendar to support long-term engagement. We also want to refine the overall experience so it feels seamless, expressive, and genuinely valuable to users.

Despite the challenges we faced, SONO showed us that music discovery can be more than efficient. It can be emotional, expressive, and personal. This project became a foundation for how we think about designing experiences that don’t just work, but feel understood.

Visit SONO today (Email Hannah2520@gmail.com for access to Spotify’s resources) at https://hanniekwak.com/sono/

To view project materials, such as our trailer, poster, design system, etc, visit http://projects.nmi.cool/2026/em/sono/