Nadia Nusatea

Dressi is an AI-powered personal stylist app for women aged 17–30. The original prototype felt technical and lacked the warmth of the founder’s mission. Which is to create a “best friend in your pocket” that celebrates unique beauty and restores joy to the daily styling ritual.

As the team leader, I spearheaded the transformation of Dressi from a utility tool into an empathetic AI styling experience. By bridging technical logic with empathy, we reduced decision fatigue and transformed wardrobe curation into a supportive journey.

Using our research and user discoveries, we formulated a problem statement to anchor our strategy. Users feel overwhelmed and indecisive because current tools offer generic suggestions and high-friction interfaces. This lack of empathy prevents confident decision-making and efficient wardrobe management. People want quick, personalized recommendations that reflect their real bodies and preferences.

To solve this, we formulated “How Might We” questions focused on occasion-based styling and wardrobe analytics. Our strategy ensures Dressi acts as a “best friend in your pocket,” turning personal data into precise recommendations that make daily dressing an effortless experience.

The Dressi brand targets younger audience while still offering an identity flexible enough to empower any woman seeking style clarity. We utilized warm, earthy tones to build a comforting foundation, paired with vibrant pink accents to inject modern energy.

Typography combines Geller with Inter for friendly human tone approachability. To ensure Dressi feels like a “supportive companion” rather than a rigid AI tool, we opted for rounded corners and soft neutral gradients for creating an inviting interface that balances empathy with confidence. So that it fits naturally into the lives of a diverse audience.

The prototype prioritizes flexibility for busy users. A “Skip for Now” onboarding flow respects short attention spans, allowing users to dive into the app immediately and complete their style profiles later. The homepage integrates a calendar for event planning alongside “Live Stories” to foster a community-driven experience.

In “My Closet,” a digital canvas lets users arrange items for virtual try-ons on personalized avatars. To remove manual friction, users can upload digital receipts for automatic tagging. Finally, “Ask Dressi” provides a natural language AI stylist that uses suggested prompts to help users build outfits effortlessly.

Testing confirmed our Information Architecture is highly intuitive, as users successfully predicted functionality across all key pages. Participants specifically praised the “Jump-In” factor, which allowed them to explore the app’s value before committing to a full setup. The Homepage and Wardrobe sections received the highest usability ratings.

While “Ask Dressi” was a favorite feature, users requested multiple outfit suggestions per prompt to increase variety. Overall, high satisfaction scores validated that the journey is logical and empowering. These success metrics confirm that Dressi successfully reduces decision fatigue while supporting a clear, user-centered path to style confidence.

Leading this project reinforced that UX for AI powered app should be built on emotional connection not just featuring technology. By shifting from sharp edges to rounded corners and selecting the warm tone & Geller typeface, we successfully transformed the user’s perception from a cold algorithm to a supportive advisor. I learned that as calling the database as “My Closet” is vital for creating a sense of ownership for our users. Merging this emotional connection with AI provides the competitive edge necessary to stand out in a saturated market.

Because the founder’s user base was unavailable, I led the team in sourcing a diverse pool of participants who prioritize self-presentation in their daily lives. If I were to lead this project again, I would prioritize iterative usability testing for the AI prompt logic during the low-fidelity stage. While our final architecture was highly intuitive, user testing feedback indicated users wanted more “creative control” over the AI’s taste. Exploring these interactive feedback loops sooner would have allowed us to architect even deeper personalization before moving into high-fidelity design.