AI-generated content was flooding the design world, but no one had figured out how to do it responsibly within a library that designers actually trusted. The questions weren't just creative — they were strategic: where could AI genuinely add value? What styles could it produce that human illustrators couldn't fill fast enough? And could it actually meet Canva's quality bar?

I led Canva's first AIGC pilot from scratch — owning the strategy, the production system, and the launch. I started with user data to identify real gaps in the library, then tested the capabilities of AI across a range of visual styles using Leonardo and Flux, exploring what each tool could do uniquely well. From there, I engineered the prompts and built custom GPTs to make the workflow seamless and transferrable — a system designed to scale beyond me, not depend on me.

Every asset went through the same rigorous QA process as anything else in the library. The result wasn't just a collection of AI-generated graphics — it was proof that AI content could meet Canva's standard, and a replicable system for how to do it again.

The collection is live on Canva.