Danny Wu leads AI product strategy at Canva, one of the world’s most popular design platforms with over 265 million users. His job consists of coordinating research, product and trust teams. to integrate AI in a way that enhances human creativity without replacing it. With a vision focused on accessibility, Wu seeks to turn complex processes into simple tasks, allowing anyone, regardless of their technical skills, to bring their ideas to life in a professional way that is consistent with their visual identity.
(MCPRO) Could you briefly describe your role as Head of AI Products at Canva and what “success” means to your team?
(Danny Wu) As Head of AI Products, my role is to coordinate our research, product, platform and trust teams to transform advances in artificial intelligence into incredible creative capabilities, allowing anyone to design anything. My focus is on integrating AI into Canva through accessible, intuitive and trusted experiences, aligned with the way people actually create. Today, AI is a superpower for some people – those with the knowledge, skills and access – and our mission is to give these superpowers to everyone, making the complex simple.
For my team, success occurs when technology takes a backseat and allows our users to unleash their imagination, putting creativity at the forefront. If people can go from an idea to a final design more quickly, maintain their brand identity with less effort, and feel more confident experimenting and working as a team, then AI is doing its job. We do this through a three-pillar approach: we make the world’s best models available to our 265 million users, we invest in Canva Research to develop our own foundational models, and we expand what’s possible through apps and integrations, including AI connectors in assistants like Claude and ChatGPT.
(MCPRO) What problem were you most determined to solve with Canva’s recent AI initiatives, and what did you learn along the way?
(Danny Wu) Our mission at Canva is to put the power of design within everyone’s reach. We see AI as an accelerator of our mission: it allows us to make design even easier and more accessible, more powerful and scalable for professionals, and advance the development of products that we could only dream of before. A good example is allowing everyone to generate and edit professional quality videos with AI, something that used to be very exclusive, expensive and inaccessible.
At Canva, we wanted to go beyond AI as a simple starting point and make it a true collaborator. Along the way, we learned that creativity is rarely linear; People constantly alternate between generating new ideas, editing, evaluating, and iterating. That vision shaped Canva AI as a comprehensive visual system that is present and useful throughout the entire workflow.
(MCPRO) How do you decide which AI features should be “magical by default” versus those that require more user control and transparency?
(Danny Wu) We design AI at Canva with intention, giving people control where it matters and removing friction where it doesn’t add value. For example, removing backgrounds: if you’re cropping a person out of an image, you just want it done, fast, and with quality. It is a technical and repetitive task where AI must intervene without friction.
Designing a complete visual piece is different. People want complete control to ensure their ideas are conveyed. That’s why, with the Canva Design Model, we built an AI system that generates fully editable designs. You can start with a simple prompt to generate a first draft and then immediately move to the Canva editor to adjust, work together and shape the result. In these cases, AI should be editable and responsive, not prescriptive.
(MCPRO) What is your product development process like when you integrate foundational models and capabilities that evolve so quickly?
(Danny Wu) Instead of relying on a single model or vendor, we have focused on building the infrastructure necessary to easily integrate a wide variety of models, including those we develop ourselves. We work closely with leading AI labs like Google and OpenAI to quickly add new capabilities, while investing in our own foundational models through Canva Research.
A key part of our process is ensuring that these models work within the context of the brand and the user’s workflow. We’ve recently expanded our AI connectors to allow you to create and refine brand-coherent designs within AI assistants like ChatGPT and Claude, with results fully aligned to anyone using it’s Canva Brand Kit. This means that AI results are no longer generic; They reflect a company’s default colors, typography, and visual style.
(MCPRO) How do you measure the impact of AI features at Canva?
(Danny Wu) We measure impact through several complementary perspectives. Productivity is key: how long does it take to go from a simple idea to a large compelling design? We look at speed, quality and creative confidence. What’s really exciting is that our foundational model now generates complete, fully editable visual designs from text instructions. This flexibility allows you to maintain control of the creative process, making it faster and more dynamic.
Our AI tools have already been used more than 24 billion times in total. It’s a clear sign of the trust and value our community finds in this technology, as well as the power of putting it at their fingertips right where they’re already working.
(MCPRO) What safeguards and governance practices are most critical to implementing generative AI responsibly?
(Danny Wu) Responsible AI is integrated from development to deployment and subsequent monitoring. We have multiple layers of protection for privacy, content moderation, and misuse prevention through Canva Shield. People who use our platform are always in control of their data with granular privacy settings. Additionally, we conduct rigorous internal reviews and testing, such as red-teaming, because we never sacrifice security for speed.
(MCPRO) In your opinion, how will AI change the craft of design in the next 2-3 years and what new skills will be the most valuable?
(Danny Wu) AI will increasingly take over repetitive and mechanical tasks, shifting the focus from craft to judgment, taste and storytelling. The most valuable skills will be the ability to define a clear creative vision, guide systems toward the right outcome, and refine results with a critical eye. We will see a more interactive and experiential design; Creators who know how to think in terms of systems and audiences will be especially well positioned.
(MCPRO) How do you envision AI agents evolving within Canva?
(Danny Wu) Agents are LLMs connected to tools that can perform specific actions and operate in an iteration loop. We are making a specific effort to make our platform agent-friendly by default. We believe there are many tasks, such as adapting content to different formats or verifying large-scale brand guides, that AI agents can handle from start to finish.
However, design is fundamentally a creative process that involves assessment and polishing. We do not imagine a future in which humans are completely eliminated; The future of creativity lies in the middle, where humans can amplify their impact by working alongside AI agents.
(MCPRO) If Canva were to offer a “team of AI agents” for a brand, what roles would they play and what conditions would need to be met?
(Danny Wu) We could have agents capable of autonomously ensuring that all content respects the brand identity, or agents who proactively suggest next steps after an initial draft. For this to work securely, agents need context (understanding the brand identity) and a robust permissions system. The results must be editable so that there is always room for human intervention. Our vision is that agents act as collaborators and not as final decision makers.
