Conscious: Connection > Consumption
Designing an AI-driven wardrobe assistant that turns fast fashion habits into sustainable style routines.
Role
UX Researcher & Sole Prototyper
Team
Solo project
Timeline
June 2024 – August 2024
Skills
Flutter, Gemini API, Python, Figma, UX Research
NOTICE: This case study’s visuals are currently being updated.
Background & Context
The rise of fast fashion has turned self-expression into overconsumption.
People are buying 60% more clothing than they did two decades ago, only to keep them for half as long. I joined UCSB’s Research Mentorship Program wanting to explore where AI and UX could shift that behavior, and reimagine how design could orient people towards wanting less, to consume less.
As an avid fashion lover, I wanted to explore how UX design solutions could shift fashion from an industry of overconsumption back to a way of self-expression.
Solution
Designing AI-driven wardrobe assistant & feedback loops
I built Conscious’s high-fidelity prototype in Figma, later connected to Gemini’s multimodal API.
01. Digital Wardrobe Upload
Users can upload images of their garments, automatically categorized and sorted by Gemini's computer vision.
02. Chat With a Personal Stylist
Users can chat with a personal stylist who focuses on reusing and building attachment to existing garments, based on the closet uploaded by the user, and can revisit past conversations.
03. A Homepage That Adapts to You
After making an account, the homepage adjusts to your shopping preferences, adapting to your overconsumption tendencies with reminders and positive reinforcement.
Problem
Fashion has been reduced to hauls…
The experience of sustainable shopping is rewarding, but it’s hard to know what’s worth keeping, what fits your style, or how to even start building a wardrobe that lasts. I wanted to design something that made that process feel good, not guilt-driven.
Research
Current platforms make discovering your style difficult…
Cognitive Overload
Managing your wardrobe, finding outfits, and tracking purchases all live in separate, disjointed tools.
Lack of Personalization
Existing recommendation systems often miss personal context — users' values, cultural preferences, or motivations for shopping.
Trend-Driven Systems
Most apps push new products instead of encouraging users to rewear what they already own.
How might we help users develop a personal, lasting relationship with their clothes—using AI and UX to make sustainability intuitive?
The Ideation Process!
Using the pain points gathered, I sketched early flows focused on reconnection—how users could rediscover what they already owned.
Sketched low-fidelity wireframes to explore flows for digital wardrobe upload, AI outfit recommendations, and style onboarding, then prompt-engineered interactions around a chat-based stylist, emphasizing quick, conversational feedback loops.
Usability Testing
Three rounds of testing with 12 participants, surveyed against 70+ people.
I conducted three rounds of usability testing to refine both flow and tone. Metrics measured: task success, error rate, and perceived helpfulness.
Average Usability Score (out of 5)
3.62
Round 1
3.95
Round 2
4.48
Round 3
User Insights
Clarify before recommending
Users preferred when the AI asked clarifying questions before suggesting outfits.
Favorites build attachment
A "Favorites" section increased attachment to clothing items.
Helpfulness resonated most with non-fashion-savvy users
Participants who self-identified as not fashion-savvy rated the app most helpful — average usability score of 4.48/5 after revisions.
Takeaways
My takeaways
01.
Simplicity is persuasive
Creative solutions can still be possible with simple, scalable designs.
02.
Conversational UX matters
The AI’s tone directly shapes user trust and satisfaction.
03.
Designing for sustainability is designing for emotion
People keep what they feel connected to.
Outcomes
What Conscious proved out
After three rounds of revisions, Conscious reached an average usability score of 4.48 / 5, with the strongest response coming from users who didn’t consider themselves fashion-savvy— exactly the audience the project set out to reach. Built and tested as part of UCSB’s Research Mentorship Program.