Unlike many AI tools that produce a final, unchangeable output, Canva's AI generates a standard, multi-layered file. This lets users treat the AI's output as a first draft that they can refine using familiar drag-and-drop tools, bridging the gap between generation and creation.

While Canva had been researching AI for years, a specific internal technical breakthrough became the catalyst for the company to "go all in." This single event prompted a rapid re-organization, pulling hundreds of people onto a centralized AI team to commercialize the new capability.

Canva avoids competing with giants like OpenAI on foundational models. Instead, it partners with them for general tasks while focusing its 100-person research team on specialized models for core design problems, like its 'Magic Layers' feature, where no adequate external solution exists.

Canva's CEO views "one-shot generation" as the first, limited phase of AI. The next frontier, or "AI 2.0," involves iterative and agentic orchestration where the AI acts as a creative partner, helping to refine a design through a series of adjustments rather than just creating a single final output.

Instead of being a monolithic model, Canva's AI works by orchestrating its entire suite of existing, specialized features like background remover. A single user prompt can trigger multiple tools in sequence to generate a complex, layered design, leveraging years of product development.

Standard file formats like .docx and .pptx are filled with complex code that LLMs struggle to parse. To build effective AI workflows, companies must create deliverables in formats that are both human-readable and AI-friendly. HTML is a prime example, as it is visually appealing for people and easily ingested by AI.

Canva views its AI as the third evolution of design interfaces. The first was pixel-based (e.g., Photoshop), the second was object-based (classic Canva), and the new era is concept-based, where users describe an idea and the AI generates an editable first draft.

Move beyond basic AI prototyping by exporting your design system into a machine-readable format like JSON. By feeding this into an AI agent, you can generate high-fidelity, on-brand components and code that engineers can use directly, dramatically accelerating the path from idea to implementation.

Canva dogfoods its AI by integrating it into core business processes. Employees can now dictate thoughts to Canva's AI, which automatically structures the input into the company's standardized "Complex Decision Making" template, complete with goals, options, and pros/cons.