POMU

AI-Powered Manufacturing Discovery Platform for Fashion Entrepreneurs

Product Design Brand Identity Design System AI/ML

Impact

350%

LinkedIn engagement growth after launch

30%

Reduction in development time

12+

User interviews conducted

Though I am no longer part of the project, my early contributions helped shape the foundation of Pomu's product and brand. The current launched product is in beta mode and still in development.

Challenge

Fashion entrepreneurs struggle to find the right manufacturer. Reddit threads, blog posts, and interviews revealed repeated friction:

Overwhelming choices without clear differentiation
Lack of transparency in pricing and capabilities
Communication barriers with overseas suppliers
No centralized platform for discovery and comparison

Pomu set out to stand apart from competitors, like Alibaba, by building trust, clarity, and speed through AI-powered discovery.

My Role

As a Product Designer, I:

Led user research with 12+ fashion entrepreneurs to understand pain points
Designed the end-to-end user experience from search to supplier match
Created Pomu's brand identity and visual language from scratch
Built a scalable design system that reduced development time by 30%
Collaborated with engineers to implement AI-powered search features

Design Goals

From the start, I set two key objectives:

1. Simplify Complex Decisions

Make manufacturer discovery feel as easy as shopping online, with smart filters and visual search capabilities.

2. Build Trust Through Transparency

Surface verified information about capabilities, certifications, and past work to help founders make confident choices.

Process

1. User Discovery

Method: Interviews with 12+ early-stage fashion entrepreneurs
Tool: Typeform, Miro
Key Insight: Founders lacked the right vocabulary, not vision. We needed to support fuzzy inputs with strong visual cues.

2. Information Architecture

We mapped a guided search-to-match journey:

Smart onboarding to understand user needs
AI-powered search with image + text input
Curated results with transparent filtering
Direct communication tools with translation support

Why it mattered: Founders often abandon platforms when overwhelmed. So we broke the journey into digestible steps.

3. Designing the Experience

We designed a dual-input search experience. I focused on:

Visual hierarchy that guides without overwhelming
Progressive disclosure of complex information
Mobile-first responsive design for on-the-go founders
Accessibility standards (WCAG 2.1 AA compliance)

4. Technical Collaboration

My background in computer science helped me to understand technical terminology and how AI algorithms work. My role involved:

Partnered with engineers to understand the CNN + NLP model pipeline, ensuring UI supported model input/output flows
Prototyped query mapping flows to visualize how image + text inputs influenced rankings
Designed edge case experiences (e.g., low confidence matches, poor image input)
Accessibility standards (WCAG 2.1 AA compliance)

AI Search Experience

Users could upload an image + describe their vision in a few words. The AI model returned visual + textual matches with confidence scores.

How it worked?

CNNs interpreted product images to identify shape, texture, and material type
NLP parsed user-provided keywords to understand manufacturing intent and constraints
Results were ranked by confidence scores, blending both model outputs

My goal was to make this feel simple and intuitive:

Clean upload interface with drag-and-drop functionality
Real-time visual feedback during processing
Smart suggestions based on uploaded images
Clear result matching with confidence scores

This was especially helpful for small business owners who had a vision but didn't always know the technical terms. By combining image and text input, we made the search smarter, more flexible—and most importantly, more human.

Design System

Together, we built a responsive design system with:

Modular component library (buttons, cards, forms)
Consistent typography scale and spacing system
Accessible color palette with proper contrast ratios
Flexible grid system for various screen sizes

Outcome: We cut dev time by 30% and ensured a cohesive, future-proof interface.

Branding

Since the platform was brand new, I had the opportunity to shape Pomu's identity from scratch. My goal was to create a visual language that felt bold, modern, and trustworthy—appealing to both early-stage fashion startups and established businesses exploring AI solutions.

Marketing

We also designed templates and sample layouts for future marketing use on LinkedIn, ensuring Pomu's brand remained cohesive across platforms—which saw engagement grow from 0 to 350+ after launch.

Reflections

Designing Pomu from scratch with another designer taught me the power of tight collaboration in zero-to-one environments. I learned to balance brand and product strategy while shipping design systems that scale—and always kept user comprehension and accessibility at the core.

Though I've since moved on, the platform's early traction and clarity-first interface remain one of my proudest contributions.