Project Overview

The AI Video Thumbnail Generator is an intelligent design app that allows creators to instantly generate, edit, and customise thumbnails using AI-powered scene detection and visual optimisation.
The goal was to simplify the tedious process of designing video thumbnails, ensuring every creator can produce professional-looking results — fast.

Timeline: 2024 (8 weeks)
Role: Lead Product Designer (UX Strategy, UI Design, Flow Optimization)
Team: 2 Designer, 1 AI Engineer, 1 Front-End Developer
Tools Used: Figma | FigJam | ChatGPT | Midjourney | Notion
Status: Completed — Prototype Delivered for Testing

The Challenge

Creators often spend too much time designing thumbnails or hire external designers — delaying upload schedules and reducing creative consistency.
The challenge: design a platform that uses AI to automate thumbnail generation while keeping room for manual creative control.

Pain Points Identified:

• Time-consuming manual design.

• Inconsistent branding across thumbnails.

• Lack of design skill among everyday creators.

The Goal

Build an intuitive web app that:

• Analyzes uploaded videos or images to identify key scenes.

• Generates thumbnails with text overlays, brand colors, and composition balance.

• Enables instant download and export for YouTube, TikTok, and Instagram.

Research & Insights

Research Methods:

• Surveyed 25 YouTubers, marketers, and short-form creators.

• Benchmarked Canva, Fotor, and Snappa AI features.

• Conducted task analysis on the thumbnail creation journey.

Key Findings:

• 70% of creators wanted more automation with control options.

• The ideal thumbnail workflow should take less than 60 seconds.

• Color contrast, face visibility, and bold typography are key engagement factors.

Personas:

• The Solo YouTuber (Michael) — wants quick, consistent thumbnails that match his brand.

• The Social Media Manager (Lydia) — manages multiple creators and needs batch generation.

Design Process

1. Ideation & Information Architecture
Outlined 3 core modules: Upload → Generate → Customize → Export.

2. Wireframing
Low-fidelity wireframes focused on clarity and reducing cognitive load.

3. Visual Design

• Adopted your Retro Future color palette for tech-modern appeal.

• Used minimal glassmorphism and glowing gradients for an AI feel.

• Created a clean, component-based UI (cards, sliders, color pickers).

4. Accessibility & Usability

• Added tooltips and hover hints for new users.

• Used WCAG 2.1 compliant contrast levels.

• Included “keyboard shortcuts” for quick edits.

5. Testing & Iteration
Prototype tested with 12 creators:

• Average thumbnail generation time: 38 seconds.

• 92% satisfaction on interface clarity.

• Added “Save Preset” feature based on feedback.

Results & Impact

• Thumbnail creation time reduced by 85%.

• Users reported 2× faster publishing workflow.

• Beta testers noted higher click-through rates (CTR) due to improved visuals.

• Developers praised the design documentation for clear component naming.

Lessons Learned

• AI features should always include “manual override” for user confidence.

• Simplicity and feedback loops make automation feel more human.

• Testing with actual creators early saves design rework.

Footer Banner

AI Video Thumbnail Generator — Empowering Every Creator with Smart Design.
View prototype on figma