MY ROLE

UX/UI Designer · Front-End Developer (React + JavaScript) · AWS Deployment Support

TIMELINE

2 weeks

PROJECT OVERVIEW

Glöd AI is a web platform that helps startups generate a short product marketing video (up to 1 minute) by uploading product photos and entering key product details such as descriptions and dimensions. The project started through our participation in the Google Gemini API Hackathon and was built by a 7-person team.

TEAM

7 (3 Full-Stack, 2 Front-End, 2 Back-End)

THE CHALLENGE

Small fashion startups and solo founders need marketing videos to stay competitive on TikTok and Instagram.

However:

• Hiring freelancers costs $2,500–$5,000 per month

• Solo founders lack editing skills and time

• Existing video editing tools feel overwhelming or “pro-level”

How might we enable non-editors to create professional-looking product marketing videos in minutes, not hours?

Design Process - Double Diamond

DISCOVER

Evaluate existing video creation tools used by small fashion startups and identify friction patterns.

Competitive Intelligence

Competitor Comparison Table

1 = Poor / High friction | 5 = Excellent / Low friction

Quantified Friction Score

Friction Index = 25 (Sum of 5 criteria) | Higher number = More friction

Key Insights

Existing video tools typically fall into two categories: professional editors that offer high creative control but come with steep learning curves and high friction, and template-based platforms that are easier to use but still require significant manual input.

Startup founders actually need:

• Zero learning curve

• A fully guided workflow

• AI-generated structure

• The speed required to launch marketing content quickly

The market offers either power or simplicity — but rarely both. GLOD AI was designed to deliver automation with clarity

I interviewed the CEO of Pink Dynamite Fashion, a mid-20s solo founder who markets on TikTok and Instagram, to understand how he currently creates marketing videos, where he gets stuck, and what he needs to post quickly without a creative team

User Interview

Interview Insights

“Editing tools are too complicated.”

“Hiring someone is too expensive.”

“I just need something clean and fast.”

Key Insights

He doesn’t want to learn editing—he wants a guided system that turns inputs into a ready-to-post draft. Current tools are either too complex or still too manual, so Glöd AI should win by generating video structure, using strong defaults, and offering lightweight edits for fast iteration

Startups need a simple, guided way to create SNS marketing videos from their product assets without learning complex editing tools or paying for expensive third-party production

Current solutions are either too complicated (professional editing software) or too costly (agencies and freelancers). What startups truly need is something clean, fast, and easy—an experience that turns basic inputs into a ready-to-post marketing video with minimal friction

Problem Statement

DEFINE

Persona

User Flow

Ethan Kim represents early-stage startup founders who need a fast, affordable, and guided way to create clean, professional marketing videos without learning complex editing tools or hiring external production support

The user flow guides startups through a clear, linear path—from entry to upload, review, generation, and export—while allowing safe backtracking and regeneration, ensuring users can create a marketing video quickly without feeling overwhelmed by complex editing decisions

This wireframe was based on user insights and competitive research: most tools are too complicated and outsourcing is too expensive, so we designed a clean, guided flow that helps users create a marketing video fast with minimal steps

Low-Fidelity Wireframes

DEVELOP

High-Fidelity

This high-fidelity prototype demonstrates Glöd AI’s end-to-end experience—from landing and pricing to guided generation, simplified editing, and export—showcasing a clean, AI-first workflow designed to help startups create professional SNS marketing videos quickly and efficiently

In the deliver phase, we launched a functional prototype and produced a marketing demo video to communicate Glöd AI’s value. Although the project did not win the hackathon, it attracted outreach from Pattern (Utah), where we presented our concept and technical approach—resulting in internship offers for two team members, with one later converting to a full-time role.

Glöd AI Marketing Video

DELIVER

Glöd AI taught me that strong UX isn’t about adding features, but removing friction. I learned that users don’t struggle with creativity—they struggle with structure and decision overload, which led me to design a guided, AI-first workflow that balances automation with simple control. If I continued the project, I would focus more on testing AI trust and refinement behavior to better understand how users iterate on generated drafts.

REFLECTION