Pyzion identifies creative intent based on work data generated from tools actually used by designers, such as Adobe Illustrator and Photoshop, and converts it into machine-readable rules called 'StyleID'. This StyleID applies brand guides in real time and generates assets that can be automatically utilized for local campaigns in over 200 brands and 200 regions around the world.
“The problem of AI misunderstanding or misinterpreting brand guidance is difficult for people, but it is an even bigger challenge for AI,” said Dominic Heinrich, head of design intelligence and technology. “Pidgeon reduces these interpretation errors and creates an environment where designers have full control.”
This system is currently in pilot phase. It is natively integrated into Adobe Creative Cloud and is powered by Firefly generative AI. Because it works within existing design tools, it can learn and adapt to designers’ workflows naturally, without any additional learning.
The key point that Coca-Cola emphasizes is that AI is not replacing designers, but rather maintaining brand consistency, automating repetitive tasks, and helping designers focus on more creative work. While existing manual brand guides were prone to errors in interpretation and application, Pizion presents a way to apply them to assets in real time, reducing time and costs.
However, industry experts point out that excessive automation can lead to the homogenization of creativity, and that AI still has limitations in understanding cultural contexts or regional differences. In addition, the structure dependent on the Adobe ecosystem is also mentioned as a limiting factor in platform scalability.
