If You Told Designers to Build Anything with AI

Toss's Design Chapter held an AI Contest last March, where designers solved problems on their own using AI. From the 122 submissions gathered over a month, they organized four key directions in which designers applied AI to real work.

First, automating repetitive tasks. They developed, together with AI, logic that automatically extracts and corrects colors for UI from images, which is now actually applied to product cards in Toss Shopping.

Second, reducing collaboration costs. They built a Slack bot trained on individual knowledge and communication data, having it draft answers to recurring design questions, thereby improving work efficiency.

Third, persuasive prototyping. Instead of static mockups, they implemented screens with working interactions using actual product code, improving the accuracy of communication with developers and POs.

Fourth, improving production quality and speed. Using AI tools like Kling, they completed high-quality motion graphics for different job roles in just a single day, even under a tight schedule.

In conclusion, rather than doing entirely new kinds of work with AI, designers are focusing on making their existing work faster, more persuasive, and of higher quality.