The quiet reinvention of a $42b business

Note: Transcripts are automatically generated from episode audio, and are not fully corrected for spelling, grammar, and formatting.

CAMERON ADAMS: We now have a quarter of a billion people who use Canva every single month. We need to make sure that Canva empowers the whole world to design through everything that we do, and that the AI we use and develop truly represents the world as well. The real power we’ve seen with AI is when you can amplify ideas and take what AI gives you and build on top of it, put your own stamp on it, make it authentic, and make it connect with the real people you want to connect with.

BOB SAFIAN: That’s Cameron Adams, Chief Product Officer and co-founder of Canva. I talked with Cam on stage at the Cannes Lions Festival about Canva’s dramatic last year as it redefined itself from a design company to an AI company. Cam challenges the idea of the so-called SaaSpocalypse, talks about how product cycles are accelerating, and argues that Canva and its peers help everyday users stay up to speed as technology shifts.

He also shares specific steps we can each take to get more impact out of AI, plus the power of exploiting what he calls each organization’s unique, unfair advantage. It’s an idea-filled romp through technology and creativity, so let’s get to it. I’m Bob Safian, and this is Rapid Response.

[THEME MUSIC]

I’m joined this morning by Cameron Adams, who is the Chief Product Officer and co-founder of Canva. Cam, thanks for doing this and inviting me to your house this morning.

ADAMS: It is amazing to be chatting with you here in the amazing cabana behind the dunes.

Copy LinkCanva’s pivot to become an AI company

SAFIAN: I was talking to Cam backstage about the challenge of this year, living dangerously with AI changing so much and Canva itself being under a different kind of pressure. Has it been a good year? What’s made it a dangerous year?

ADAMS: I’d probably say it’s been an interesting year. I think we’ve always been running a marathon at Canva. We’ve always had an incredibly long-term vision. This year feels like we’re sprinting that marathon just a little bit.

SAFIAN: You said that you now consider Canva to be an AI company that does design rather than a design company that does AI. That seems like a big change to make in a company as big as Canva.

ADAMS: It has been a big change. Having a strong bench of AI researchers and ML engineers has enabled us to move rapidly on the product and, particularly, to develop our own unique blend of AI. It’s not just repackaging an LLM. It’s deep technical research and unique IP that now exists within the Canva platform.

The last year, the team has scaled up, which has enabled us to ship more, but we’ve also entered a new cadence of product shipping. To me, it feels really exciting because it’s almost a return to our startup roots. It’s not thinking about 6,000 people and moving them all like the Titanic. It is literally a bunch of speedboats now going around and shipping as much customer value as we possibly can on a week-by-week cadence.

SAFIAN: When you’re moving at the pace of AI, I guess your product pace has to be faster than it was. How much do you worry about what that means to your users? If the product changes too fast for them, they’re not looking for it to change. They’re looking for it to continue to do the things they want it to do. Is all the change behind the scenes, or how do you get the customer up to speed with you?

ADAMS: Definitely, there’s a lot of behind-the-scenes change in terms of how quickly our teams are putting features out, and particularly how quickly we get it out to our own internal team. Every great software company, every great AI company, is shipping to itself first. So we’re literally shipping new stuff to the Canva team every day to try out, to see if it’s good-quality enough, and to see how we need to change the experience to map to it.

But as you say, if you did that to your customers, they would be incredibly frustrated because they just want to turn up and get their job done. So you have to be careful when you roll out a very major change to app architecture or anything like that, but you can roll out smaller stuff quite easily. You can roll out improvements to models, a new filter that you could apply to an image, or a new kind of workflow that you can apply to your images. So we try not to move all the pieces around too much, but adding in new pieces you can do fairly scalably.

SAFIAN: When you think about the playbook of how you’ve had to refocus the company and the product, do you look at some of the companies in Silicon Valley and say, “Oh, we’re going to do it that way,” or, “We’re not going to do it that way”? You guys are far from Silicon Valley in Sydney.

ADAMS: We’ve made, I think, a couple of really pivotal changes to the way we work with our teams to enable that. First, it is: use whatever tool you need. So we’re not mandating Claude or ChatGPT or Gemini, or whatever tool you want to use in your part of the company. You can figure it out. You’ve got budget to try out tools, figure out your workflow and your process, and tackle the problems that you have. Giving people that freedom is really important, because they need to feel comfortable experimenting. If you force a tool on them, they’re just going to do that very begrudgingly, and they’re not going to enter into this very experimental mindset that we need them to.

The second part is giving them the time to explore and try things out. I’m sure we’re all terribly busy and trying to get 101 things done every single week, and you naturally fall into the patterns you would normally do things in. You’ve got your tried-and-trusted tools, you know how they work, and you’ll always reach for them. So you’re not looking for that medium-term and long-term improvement that you can get through some of these tools. So we give our staff plenty of time to sit back, put down their usual tools, step out of business as usual, and experiment.

Something we did a month ago called AI Discovery Week really speaks to that. For the whole week, we said to people, “Please don’t do your normal work. We want you to think of the problems that you have, the tools that you’ve heard about, the opportunities that you’ve heard from colleagues in other industries, and we want you to try those out for the entire week.”

At the end of the week, we did a massive show-and-tell where people showed off what they did, and we had about 400 or 500 different projects that people showed off, employing AI in different ways, from our legal team to our finance team to, of course, our product and engineering teams. It was super energizing.

SAFIAN: Are those tools that you’re actually implementing, or are they more like prototypes for things?

ADAMS: I think almost half of them are now pretty much in production, either in production internally or in production actually going to our customers. So our legal team rolled out a trademark-checking application, because they spend thousands of hours every year thinking about our product names, what we’re shipping, and whether, all across the world, we can trademark that and legally use that. They’ve now rolled out a tool that has literally saved them 1,500 hours of work.

Copy LinkUsing AI tools to solve real problems

SAFIAN: How do you think about being disciplined about experimenting with, growing, and using these tools, but not overdoing it? I know it’s supposed to make me more efficient, but sometimes I find I’m doing so many different new things that maybe I’m not actually being more efficient.

ADAMS: Personally, I don’t like just picking up a tool with no real use case and experimenting with it. I like to take something that I’m working on now or a problem that I’m feeling, and then bring a tool to bear on that and learn through that process. That’s how we’ve seen our staff also get the most value out of AI experimentation: when they’re dealing with a real problem.

SAFIAN: So you launched Canva AI 2.0 in April, which is a conversational assistant that helps generate designs from descriptions rather than from templates. How do you know when you’re amplifying creativity versus automating it away? That’s some of the anxiety that comes with using AI tools.

ADAMS: One of the principles that we adhere to as we’re developing an AI product is that we need to keep humans in the driver’s seat. We don’t want to create a product that automates everything away, that automates creativity and taste and judgment, and purports to give you a solution where you don’t need to interact with it at all. The real power we’ve seen with AI is when you can amplify ideas and take what AI gives you and build on top of it.

We’ve seen many dead ends in other products where you ask it for something, it gives you something back, and you can’t do anything with that. You either take it or leave it. It’s a flat image, or it’s a swath of text that you then have to copy and paste somewhere else. Canva is all about enhancing people’s creativity and getting them involved in that creative process.

So yes, AI can give you 101 generated images. It can generate copy lines for you, or it can help you draft a blog post, but you need to take that, put your own stamp on it, make it authentic, and make it connect with the real people you want to connect with. So every part of AI in Canva is editable. Anything that you produce, you need to be able to take into a normal Canva design, share with your team, collaborate with your marketing folks and your sales folks and your HR team, and bring them into it as a first-class citizen, not as a tack-on to technology.

SAFIAN: I always felt like that was part of Canva’s promise: it helped me, as a user, have better taste. With AI, you’re sort of providing that originality and maintaining soul at the same time. Can you teach an AI model to have empathy, to understand the emotional reaction that humans are going to have?

ADAMS: I’m not sure whether you can teach it empathy, but you can give it some understanding of the context we’re living in and the goals we’re trying to achieve. That’s very much what we do when we’re integrating this AI into our platform. When you talk about conversational AI, it literally is designing a conversation, which is kind of a new field for many different people. But designing a conversational interface is quite different from something that you point and click.

SAFIAN: Right, which is the way I used to engage with Canva.

ADAMS: Exactly. You need to put a lot of thought into it, and there are new ways you can shape that conversation you’re having with an AI. There’s a ton that goes into the system prompt. There’s a ton of orchestration between different agents, and all of that forms part of the experience. So it’s definitely a frontier for building product, but one our team has been really excited about entering. And when you see a well-designed conversation that gives you insight into the creative process, like an AI that helps you understand better font choices or why it chose this color so that you can roll that into your own taste and judgment, I think that is the type of conversation we want to be having with AI rather than just a mercenary or functional interaction.

SAFIAN: When you’re tuning the models for your use, you’re putting guardrails in to say, “Don’t put these things together.” There are certain design principles baked in behind it.

ADAMS: We have an amazing understanding of design, both from an internal expert level, whose knowledge and experience we can bake into the AI experience. We also have a tremendous amount of data from our customers as they’ve interacted with designs. That can help the AI understand what people want to achieve. So we’ve baked all that into what we call the Canva Design Model, which is actually the world’s first foundational design model that we released last year. It goes far further than many of the general-purpose models you have access to, and it specifically knows what good design looks like and what you are trying to achieve with your design.

SAFIAN: I want to pause here to underscore what Cam is saying. Canva’s AI platform is bespoke to its users, to the craft of design. In that way, it’s like a window into how AI is evolving to be more purpose- and project-specific. So is there a competitive advantage in working across multiple AI models like Gemini and ChatGPT and the like? And what’s it been like to help lead Canva through the storm of the so-called SaaSpocalypse? We’ll talk about that and more after the break. Stay with us.

[AD BREAK]

Before the break, Canva’s Cameron Adams talked about a year of dramatic change from being design-first to being AI-first. Now he explains why he’s not phased by the so-called SaaSpocalypse, what Canva needs in its pursuit of a billion users, and the power of exploiting each organization’s unique, unfair advantage. Let’s dive back in.

Copy LinkCan Canva survive the SaaSpocalypse?

There’s that phrase that became popular earlier this year: SaaSpocalypse.

ADAMS: Yeah.

SAFIAN: The idea was that businesses based around software, not unlike Canva, were suddenly vulnerable to obsolescence because of what the models could do. Is that something that worried you or spurred you in a different way?

ADAMS: I don’t think it worried us, but it certainly gave us an interesting viewpoint to take as we thought more deeply about building an AI platform. I think this notion that there’s going to be one place the entire world goes to do all of its work is not going to be reality. Tools like Canva that focus on particular areas, like visual communication and visual content creation, will be able to create an experience that is far deeper, far richer, and gets you better results. At the end of the day, that’s what we all want. We want better results for what we’re trying to do, and Canva can certainly do that with visual AI.

SAFIAN: There’s a lot of talk these days about the responsibility of AI companies, that just because you can do something maybe doesn’t mean you should. What kind of responsibility do you feel Canva has?

ADAMS: We feel a very deep responsibility. We now have a quarter of a billion people who use Canva every single month. Being able to fuel more of that through AI, through better product development, is an incredible responsibility that we feel. We are very optimistic that Canva will get to over a billion people using Canva itself. That puts a responsibility on us to make sure every person in the world can access that, not just the rich few, not just a few people who can speak English, but literally everyone in the world who speaks different languages, who has different economic capabilities, and who operates on different devices. We need to make sure Canva empowers the whole world to design through everything we do, and that the AI we use and develop truly represents the world as well.

Copy LinkHow creators can adapt to AI

SAFIAN: I will say that when I speak publicly at this event on creativity that we’re at here, people talk about all the great things AI can do. More privately, people worry about their creativity being taken away or how AI is going to step in and take jobs. Do you have advice for people who want to stay creative but also don’t want to fall behind technologically? Can we just trust that the Canvas of the world are going to take care of that for us? Or do we have to be learning these tools and figuring out how to use them in addition to that?

ADAMS: You can definitely trust the Canvs of the world to enable your creativity, but you also need to take a hand in this moment and in the change that’s happening. I can fully understand why a lot of people feel a bit fearful or hesitant about what’s going on, because that always happens in moments of change.

Speaking personally, I’ve been through a tremendous amount of change in my life, and embracing that change has, without a doubt, been the most important part of shaping a better life for myself. In this moment of AI, looking at the tools and figuring out how it integrates into your creative process is really important, because no one in this room has the same creative process. Some of us think in images, some of us think in words, some of us sketch, and some of us work entirely on a computer.

That is a legitimate creative process. Your creative process is your own creative process. Figuring out how AI fits into that is the most important thing you can be doing right now. That doesn’t mean you need to use AI for everything. It means you can use AI to convert your sketch into 15 other variations, or you can brainstorm on a whiteboard about what an event space looks like, or you can use it to give you the first draft of a marketing email.

SAFIAN: And you don’t have to embrace a conversational interface if you don’t want to.

ADAMS: It doesn’t have to be conversational. One of the things we focus on in Canva is that not everything needs to be spoken or written. You need to be able to work visually with your tools as well.

Copy LinkWhy Canva stays model agnostic

SAFIAN: You talk about all these new things Canva can do now. It’s still early days with AI. How far along are we? How much do you feel Canva is going to change from here?

ADAMS: We are just getting started. It’s actually a phrase we have inside Canva that we’re only ever 1 percent of the way there, because the other 99 percent just keeps growing and evolving and changing. It’s honestly exciting, because that means there’s still an incredible amount of opportunity ahead for us.

SAFIAN: Do you have a favorite model, or are you model agnostic?

ADAMS: We’re pretty model agnostic. There’s constantly leapfrogging all the time.

SAFIAN: Yeah.

ADAMS: It’s great to have both Gemini, Claude, and ChatGPT duking it out, because that’s what drives innovation. Innovation is never driven by sitting on your laurels and thinking the world is fine. Innovation is driven by people pushing you. It’s great to have folks like that pushing us, and us pushing them. It creates, I think, a better world for users because they get more choice, more innovation, and better things out of it.

SAFIAN: One of your colleagues said to me, “Every company needs to understand its own unfair advantage in this platform shift.” What defines an unfair advantage? What’s Canva’s unfair advantage?

ADAMS: I think our unfair advantage is being able to tap into people’s creativity at scale. We can ship incredibly impactful features at the flip of a switch. That goes out to a huge audience, who then immediately take that and start deploying it on real projects all over the world.

We’ve got over 500 designs created in Canva every single second, and every one of those designs has a story behind it. It has a story of someone trying to build their company, or the story of someone trying to teach a better class in a high school down the road. I feel very privileged to see those stories coming through the Canva platform. We use them to inspire our users with the stories of other design creators, and we use them to improve the product as well.

SAFIAN: For the folks who are here listening, they may be creators, they may be running or building their own businesses or their own products. How do you think about your business differently than you did a year ago? And what advice do you have for them on what they should be focusing on in the year ahead?

ADAMS: Probably the way I think about business differently is that we’re now closer to the goal end of what people are trying to achieve rather than the design end. A lot of people used to come to Canva just to get a picture out of it. They wanted something for their social media channel, or they wanted a flyer they could send out to their customers. Now we’re getting closer to the goal. The goal is finding more customers, increasing product sales, having more impact through my nonprofit, or finding more volunteers.

With AI in that creative workflow, we can now get closer to those goals because people can come to us and ask, “How do I get more customers? Given my product, given the photos I took yesterday, given some video footage I took at this conference, what can I create that is going to get me closer to my customers?” And that’s a really exciting question to be able to ask inside a product.

SAFIAN: Cam, this has been great. Thanks again for inviting me to your cabana.

ADAMS: No worries. Thank you.

SAFIAN: AI’s role in moving Canva users from being image-focused or design-focused to being goals-focused is a pretty radical step, and one not everyone is going to be comfortable with. But it does offer the promise of helping us be more sophisticated and strategic in our decision-making.

I was struck by Cam’s assertion that Canva is still only at 1 percent of its potential. That’s both terrifying and reassuring. I mean, if a $42 billion company is only 1 percent evolved, the future ahead of us is pretty murky. But on the other hand, moonshot aspirations are what lead to progress and opportunity. Plus, they’re fun. So here’s to practicing gratitude for the infinite starting line. Even in an AI world, we’ve got to take some creative license. I’m Bob Safian. Thanks for listening.