LOS ANGELES: The new Enhanced Voice tool in Magic Studio, which removes background noise from the video files you may be editing. An individual’s voice, which is more of a writing style, which Magic Write can use to personalise a document or piece of text. AI powered style matching if you’re working on a series of presentations. These are just some of the tools that Canva, popular suite for visual communication and workplace functionality, is adding as part of a big set of new updates which will allow it to become more relevant for enterprises, as well as use cases that involve teams required to collaborate on content creation as well as domain and specific functions within businesses.
Speaking exclusively with HT on the sideliners of the company’s first global Canva Create keynote in Los Angeles, Cameron Adams, co-founder and Chief Product Officer, makes it a point to note India’s importance for Canva. “India is actually our biggest market. It’s huge and one of our fastest growing as well. Our user base there more than doubled in 2023,” he says. The platform now has more than 180 million monthly active users, including 20 million paying premium ‘Pro’ subscribers, clocking $2.3 billion in annual revenue. Individual users remain at the core of its strategy, and Adams insists that isn’t about to change. Instead, enterprise and workplace focus with the new Canva Enterprise and teams’ tools, is an additive to the user demographic, and not exactly a pivot.
Unlock exclusive access to the latest news on India’s general elections, only on the HT App. Download Now! Download Now!
Adams speaks with HT about competing with big tech companies including Adobe and Google, AI model development in-house, their approach to data privacy, plans for AI content identifiers and more. Edited Excerpts.
Q. Canva’s versatile, such as the Magic Studio, an AI marketplace, Docs and now a number of workplace tools. Does that put a big target on Canva’s back, since you’re competing with apps from a number of tech and AI companies?
Cameron Adams: To be honest, we haven’t thought about it that much. We just constantly want to build a great product and we really love building new products, what we have and listening to customers about what they like and what they want to see next What we’re constantly thinking about is the road ahead and how to truly empower the world to design.
It’s a pretty huge mission. The last 12 years of Canva is still a drop in the ocean of what we want to build over the next decade with our own vision of what we want to build and incorporating feedback. The new enterprise features, for instance, a large part of the development of that has been from feedback from enterprise customers. It wasn’t something that we focused on a huge amount three or four years ago because That’s what wasn’t where our heads were, but as Canva has grown inside enterprises, it is often quite organically.
We have over 90% of the Fortune 500 companies using it now. And as it has grown there, we’ve gotten increasing amounts of feedback. So, the development of Canva Enterprise, which we’re launching today, has been driven by those set of requirements as well as our vision of how we would love to see collaboration happen. Throughout our history, it’s never been about just listening to the customer, and it’s never been just about us forging ahead and doing whatever we want. It’s really about merging the two together to create something that’s innovative, that people can appreciate and immediately see how they would use it.
Q. Is there pressure to speed up R&D, not only to keep products relevant, but also to keep pace with competition?
CA: We believe it’ll be incredibly important to not leave any of our existing customers based behind, and a lot of the new features we’ve launched, will still really appeal to students and teachers and individual business owners. But we are seeing rapid growth inside enterprises and teams, which I think is because more and more people are getting comfortable collaborating online. It’s an expectation that your designs immediately sync across all your devices, and can immediately see the comment from someone else in your office, on the doc.
That really fuels the spread of the product as well. Today, Canva’s been amazingly driven by word of mouth. So, being able to consolidate that and then have various teams grow Canva in a way that they’re comfortable with inside their organization. That is going to be a great channel for us. But it’s really additive, and not a pivot for us.
Q. How are you positioning the enterprise functionality, tools for teams and an updated AI layer for Magic Studio? Does it necessitate a change in messaging to consumers and enterprises?
CA: Roughly half of the employees at Canva are engineers, which I think is quite an unusual split particularly for a company of our size. Alongside, we have a big team of product managers and product designers, as we’re a very product-led company. As a result, we’re always thinking about how we can improve it. What new technology is available, how can we adapt that technology into an experience that works for our customers? AI has been a great example of that. We were very fortunate to have laid the foundations very early around six or seven years ago with a machine learning team and that meant that as generative AI arrived on to the scene, we were rapidly able to capitalize on it.
It actually took us six weeks to launch our very first generative AI product, which was text to image. It was based off of a small hackathon project that one of our engineers did and it was so impressive, we decided let’s get this in front of 85 million people at the time. That six-week period is probably one of the fastest products we’ve ever delivered. But you can only do that when you’ve got the muscles built up, when you spent the last 10 years training and building the team and just being flexible, adaptable and agile. R&D is hugely important, and we’ve actually just brought on a new chief scientist who is keenly focused on AI as well as a bunch of other interesting technological areas inside Canva. We will have an AI research and development roadmap for the next several years.
Q. Is there a plan to place content credentials or identifiers of some sort with all generations done using AI?
CA: We definitely want people to be able to transparently know when AI has been used to produce content. We currently do put metadata that enables you to know when an image has been AI generated. There’s still a bit of work happening internationally and between different organizations to figure out the right standard for doing that. We’re actually the only Australian company to be part of a US Consortium (editor’s note: Adams is referring to the US AI Safety Institute Consortium, announced in February this year; Adobe, Apple, Cisco, Meta, Microsoft, HP, IBM, OpenAI and Stability AI are also part of) that’s working with the government. A part of that conversation is about AI regulation policy, and specifics like watermarking and content detection. We’re really keen to have robust tools so that people can know the veracity of an image, and whether it’s entirely human or whether AI played a part.
Q. What is Canva’s approach to use data and privacy, more as AI figures prominently?
CA: We think it’s really important for users to be able to understand where their data is going and to be able to control that. We think it’s really important for users to be able to understand where their data is going and to be able to control that. We launched canvas Shield at the end of last year and it covers AI as well as a bunch of other privacy and data settings. We have given people easy to understand switches in their settings that they can switch on and off if they want to include their data in any AI training that we might do, or they can say, no, I don’t want that.
Q. This summer, we’ve already seen AI take significant steps forward with OpenAI and then Google. How do you foresee AI evolving over the next few years?
CA: It’s moving rapidly, which is exciting. We see the future of AI is really weaving its way into your workflow. Rather than just being this one tool that you go to et an image or when you need, some text generated. The landscape at the moment is very fragmented. If you do want to go to an image generation, you might go to one tool. If you want some text, you might go to another tool. If you wanted anything to work on your code, you might go to a third app. This just creates all these friction points where you have to upload stuff, download stuff, you have to copy and paste and a lot to figure out. It’s incredibly inefficient.
It is also causing Enterprises a lot of headaches. We did a survey with CIOs around six months ago and they were really keen to introduce AI into their workflows. In fact, nine out of 10 Indian CIOs said that they were interested in doing that, but seven out of ten of them said they were really suffering from too many apps inside the organization. That means there’s too much to manage and administer, and too much cost from having to pay for all these different applications. They really want to consolidate.
We want to counteract that fragment action, app sprawl, and help people in using the software in a better way. A way that’s more intuitive to them, that doesn’t require them to bounce around between apps and is more efficient. By bringing Magic Write as a text tool into Canva, with Magic Media as an image and illustration generation tool and Magic Switch that can convert a whiteboard into a document, that’s all part of the one platform that’s seamless, and a workflow of going from idea to reality
And it’s a lot cheaper, frankly. Paying $15 or so a month for Canva as opposed to $10 a month for each of a whole bunch of other tools just makes it a lot cheaper and easier to roll out. That’s the way we’re approaching it.
Q. How important is the role of partners, particularly with the AI marketplace as it develops?
CA: It’s been really great to partner with the world’s best AI providers, such as OpenAI and Runway ai, who help us with the text to video feature. They’ve been amazing to work with. You know, we get a lot of of the OpenAI’s tech well before it hits mainstream, and it’s really cool to see how fast they’re going and how willing they are to partner with us. We were one of the very first companies to write a GPT for their ChatGPT program and we’re among the top places for most usages of the GPT.
It’s been really good to be working directly with partners and then we’ve spent a lot of time with our ecosystem building an API that’s really attractive to developers who want to access the 185 million Canva user base.
We have got over 100 apps in the Marketplace now, and a lot of those are AI apps. AI developers have been really keen to get their apps on Canva. AI generators, even AI presenters that you can put on a presentation or adding music to your presentation which is generated using AI. A whole slew of different AI use cases and these neatly fills the cracks, because we spend a lot of time building our own AI models. Inside Canva, some of the core things in the experience, we have partners such as OpenAI to handle the medium-sized rocks and then all the little gaps in between can be filled by the developers who want to bring their knowledge to a particular area.
Q. So how much of Canvas’ AI layer relies on in-house developed models, and how dependent are you on OpenAI and other AI companies?
CA: It’ll be difficult to calculate a percentage split, but I can say a huge amount of the AI in Canva is developed in-house. We’ve got a got a huge team of machine learning engineers, which is close to 100 machine learning engineers within Canva, and they’re constantly working on our own tech. Companies such as OpenAI are amazing at building large language models. Frankly, we like to drill in where we think we have the strongest know-how, the most data and the best understanding of what needs to go in. Basically, anything that’s related to the visual side of things, we really focus on building in-house. That things like Magic Design and Magic Video which we’ve launched for the enhanced audio. Developing those things in-house is because we see the value. Simply put, we don’t need to build everything ourselves. We use the best where we can get it, and when we think we’re the best, we’ll build it.