ensemble - designing an app to make getting dressed easier and more intentional.
TIMELINE
Winter 2025
(8 Weeks)
ROLE
End-to-end Designer
TEAM
Independent Project
w/ Peer Feedback
SKILLS
UX Research
Prototyping
Visual Design
AI Integration
getting dressed is more than a routine; it’s a daily act of self-expression.
Unfortunately, in the rush of everyday life, it can also be time-consuming and lead to decision fatigue. Motivated by my love for fashion and my interest in user-centered design, this project explores how digital tools can support personal style and daily decision-making. I led the end-to-end process, from early research to interface design, to create an experience that feels elegant, helpful, and empowering.
SOLUTION HIGHLIGHTS
the solution: a wardrobe tracker with an AI assistant to help create outfits
assemble: an AI styling assistant
Use natural language to describe your needs, and assemble will create an outfit based on the weather, the events, and the user's preferences.
RESEARCH
why is it difficult to get dressed?
We started by conducting one-on-one interviews with 5 users about their wardrobe experiences and clothing habits. These conversations revealed some common challenges:
COMPETITIVE ANALYSIS
To understand how Ensemble compares to existing solutions, I analyzed several popular apps with similar goals. Each offers its own strengths and areas where it could improve.
APP NAME
FOCUS AREA
STRENGTHS
LIMITATIONS
Indyx
Outfit planning
+ styling services
Offers personalized styling by experts
Styling services
are high-cost
Whering
Wardrobe organization
+ analytics
Detailed stats
+ organization
Overwhelming data presentation
ACloset
Style inspiration
+ community
Inspiring looks
+ ideas
Inspiration isn't tailored to user
Stylebook
Closet cataloging
+ planning
Comprehensive closet catalog
Very bare-bones
Cladwell
Capsule wardrobe
+ daily outfits
Focused guidance
using AI
AI doesn't suggest diverse range of outfits
USER PERSONA
To design a solution that fits users' lives, it's important to understand their daily challenges, goals, and routines. I created a user persona to better understand a typical user.
Vivi
she/her
BACKGROUND
Vivi, 28, is a marketing professional with a busy schedule of work and social events. She enjoys fashion and wants her outfits to feel polished and confident, but mornings are rushed and she rarely has time to plan. Quick closet decisions often leave her feeling uninspired.
FRUSTRATIONS
Getting dressed feels stressful and eats into her limited time. Without a clear system, she defaults to quick choices that don’t always fit the day’s events. She wants a smoother morning routine — one where she already knows what to wear, feels confident, and saves energy for the rest of her day.
IDEATION + DESIGN
version one incoming!
The goal was to reduce the friction of getting dressed by offering a place where users could catalog what they own, plan ahead, and reflect on their outfit choices over time.
V1 FEATURES
USER FEEDBACK
unfortunately, users still aren't able to quickly get dressed.
User feedback revealed that even with a well-organized wardrobe, decision-making remained a major challenge. Many users still found it difficult to come up with outfit ideas, especially when pressed for time. These insights highlighted a clear next step: find a way to offer more active support during the getting-ready process.
SOLVING DECISION FATIGUE
if organization isn't the answer, then what is?
Ensemble was helping users manage their wardrobe, but it wasn't helping them make decisions. I needed to shift from passive organization to active support.
Before settling on a solution, I explored a few different ways the app could help users move from "I don't know what to wear" to "I know exactly what I'm wearing."
ASSEMBLE
meet assemble: your ai styling assistant
After exploring these approaches, we realized what users actually needed: a flexible, context-aware partner that could meet them wherever they were in the decision-making process.
Some days, users know exactly what vibe they want: "Something professional but approachable for a client lunch." Other days, they're starting from scratch: "I don't know, just help."
A conversational AI assistant could handle both:
ASSEMBLE FEATURES
Natural
Mirrors how people think about getting dressed (through conversation and refinement)
Contextually Aware
Assemble considers the weather, calendar events, and past preferences simultaneously
Collaborative
Users can accept whole outfits, swap individual pieces, or revise as needed
Flexible Inputs
Users can be as specific ("smart casual for a team lunch") or vague ("something cozy") as they want
Unlike the other approaches, Assemble doesn't force users into a rigid system. It adapts to how they want to interact with it, whether that's a quick suggestion or a back-and-forth styling session.
assemble in action!
Users talk to Assemble the way they'd talk to a friend helping them get dressed. They can be specific ("professional outfit for client meeting") or vague ("something comfortable"), and Assemble figures out the rest based on context.
THE BASIC ASSEMBLE FLOW
If the first suggestion misses the mark, users can type a new prompt with more specific direction. Assemble generates a fresh suggestion based on the updated input. The process repeats until users land on something they feel confident wearing.
USING ASSEMBLE TO MAKE CHANGES TO AN OUTFIT
how assemble thinks
Assemble's job is to make outfit decisions easier, not harder. That means suggestions need to be practical, personal, and actually relevant to what you're doing that day. To pull that off, Assemble looks at four main things: weather, your calendar, your style preferences, and your recent outfit history. Here's how each one shapes what you see.
WEATHER CONTEXT
Assemble checks current and forecasted weather to suggest temperature and season-appropriate clothing—temperature, precipitation, wind, and time of day all factor in.
EXAMPLE:
It's 52°F and drizzling? Assemble prioritizes layers and water-resistant pieces over the sundress you wore yesterday when it was sunny.
CALENDAR EVENTS
Your schedule shapes what you wear. Assemble looks at event types, locations, duration, and timing to match outfits to your day.
EXAMPLE:
"Client Presentation" at 2pm followed by "Drinks with Sarah" at 6pm? Assemble suggests something business-casual that transitions easily—like a blazer you can remove between events.
STYLE PREFERENCES
Assemble learns what you actually like to wear by tracking outfit ratings, wear frequency, and style patterns.
EXAMPLE:
If you consistently rate skinny jeans as "not for me," Assemble stops suggesting them, even if they're in your wardrobe. Ownership doesn't equal preference.
PAST OUTFITS
Assemble remembers what you've worn recently to avoid repetition and encourage variety across your wardrobe.
EXAMPLE:
Wore your black turtleneck three times this week? Assemble will deprioritize it unless you specifically ask for it, helping you explore beyond your "safe" pieces.
TRANSPARENCY, TRUST, + ETHICS
styling notes: the "why" behind every outfit
Every outfit suggestion comes with styling notes that explain the rationale behind it. When AI makes decisions without explanation, it positions itself as an authority you're supposed to trust blindly. That's especially problematic in fashion, where there's no objective "right answer" and personal expression matters deeply. By showing its work, Assemble acknowledges that it's offering one perspective, not the truth.
Styling notes also prevent the AI from becoming a crutch. The goal isn't to make users dependent on Assemble forever; it's to build their confidence and understanding over time. When users see why certain colors complement each other or why layering works for fluctuating temperatures, they internalize those principles.
giving advice, not orders
When users select outfits that might be impractical given the context, Assemble gently points it out, but it always makes it clear the final decision is the user's. The AI focuses on comfort and functionality, not style judgment.
Here are some examples of practical feedback:
Assemble provides information users might not have considered (weather conditions, terrain, event logistics) but never presumes to know what matters most to them. Maybe the heels are worth the discomfort for a special occasion. Maybe they're willing to risk the white pants. The AI's job is to surface practical considerations, not make decisions. Users can ignore suggestions entirely, and if they consistently do, Assemble learns to stop flagging those concerns.
ASSEMBLE SUGGESTS CHANGING SHOES
fashion is very personal
There's no objectively "right" outfit, and what works for one person might feel completely wrong for another. I designed Assemble to make suggestions without imposing taste. It won't tell you "this looks bad" or rank outfits as "better" or "worse." Instead, it offers options with explanations, allowing users to develop their own point of view. The goal isn't to make everyone dress the same way; it's to make personal style more accessible and less mentally taxing.
OUTCOMES
did ensemble help people get dressed?
With Assemble, the app began to address that missing layer: inspiration. By introducing AI-powered outfit suggestions tailored to weather, events, and personal preferences, the app started to offer real support in users’ daily routines. Instead of asking, “What should I wear?” and having to manually sift through their closet, users could now receive thoughtful starting points — and tweak them as needed.
This shift not only helped reduce decision fatigue but also made the experience feel more dynamic and empowering. The AI didn’t replace the user’s sense of style — it encouraged it. It turned the app into a collaborative space where users could explore new combinations, feel more confident in their choices, and ultimately get dressed with more ease and enjoyment.
TIME SAVED
Users spent less time agonizing over their outfits.
SMOOTHER MORNINGS
Users described mornings as less stressful and more effortless.
ENGAGEMENT WITH CLOSET
Weekly outfit logging helped users rediscover forgotten clothes.
CONFIDENCE BOOST
Users felt more polished and self-assured in their outfits
REFLECTIONS
what did i learn from this project?
Working on this project reminded me that good design isn’t just about creating a clean interface or organizing information; it’s about solving the right problem. It’s important to step back and make sure to fully address the initial issue users were facing. Users didn’t just want structure; they wanted support.
AI is a useful tool, but it is important to ensure that it is transparent with its reasoning and to have room for the user to express their personal style. The most important features aren’t always the ones that are the most complex; rather, they’re the ones that understand the user’s mental load and offer relief in the moments it matters most.
