An orchestra, where everything works seamlessly. Canva co-founder and CEO Melanie Perkins credits one of the company’s engineers for coining this analogy to describe the new direction the 13-year-old platform is taking. The focus is now on artificial intelligence (AI) agents, which will take guidance from humans to execute complex tasks. Adobe also, earlier today, announced a new Creative Agent on its platform.
Agent AI is on the agenda for tech companies, although each company is approaching the conversation differently. Creative platforms are looking at AI as a collaborative element in designing, a subtle change from the theme repeatedly set by AI companies including OpenAI, Nvidia, and Anthropic, which focuses on replacing humans in the creative workflow.
Perkins is confident that the combination of agentic orchestration, conversational design and object-based intelligence in Canva AI 2.0, announced today at the Canva Create keynote, will become a “true creative partner” with “the tools to transform ideas into fulfilling results.” An interactive prompt can not only generate multiple elements or types of content, but can also help with editing.
The foundation for this agentic AI integration began to be phased in last year, with the release of Canva Code, followed by a Creative OS that refined AI in docs, photo editors, and video editing. Canva hopes the new agentic tools will give them a better handle on business and team workflows.
It’s competing not only with Adobe’s Creative Suite, but also Apple’s Creator Studio bundle, as well as a much broader canvas including working platforms. This month’s acquisition of Australian AI companies SimTheory and Orto adds content to an intensified agentic workspace focus. The keynote comes in on strong momentum with 265 million monthly active users, of which 31 million are paid subscribers, and $4 billion in annual revenue.
Adobe’s new Firefly AI Assistant can organize and execute complex, multi-step workflows in Firefly, Photoshop, Premiere, Lightroom, Express, and Illustrator. David Wadhwani, president of the creativity and productivity business at Adobe, says this is “a new era of agentic creativity, where you direct how your work takes shape and your vision, voice and taste become the most powerful creative tools”.
The company says the Firefly AI assistant will be available in public beta in the coming weeks.
US venture capital firm Andreessen Horowitz (A16Z) says Canva is behind only OpenAI’s ChatGPT and Google’s Gemini in the global AI usage hierarchy, and ahead of DeepSeek, Grok and Cloud. Additionally, a16z and market researcher YpitData noted that Canva recorded a 101% year-over-year increase in customer spend, ahead of Replit (78%), Vercel (72%), HubSpot (63%), Box (60%) and Figma (47%).
Adobe is leveraging the Foundry architecture, giving companies the freedom to train Adobe’s Firefly models with their guidelines and data in a commercially secure structure. Nvidia’s Agent Toolkit tests complex workflow loops, including whether an AI agent can iterate on a design to match defined aesthetic guidelines.
“These aren’t the chatbots of yesteryear giving generic suggestions. Our AI assistants put context into action, understand your creative intent, anticipate your needs and learn from what you choose to share, adapting to your style and getting better over time,” explains Eli Greenfield, Adobe’s chief technology officer and SVP of Creative Products.
Collaborative design tool Figma has introduced an Anima agent, often referred to as “God Mode”, which can automate UI design, and even convert finished designs into production-ready code.
Affordable model, and the right connectors
“We have a saying now that Canva is more AI, less UI,” says Cliff Obrecht, Canva co-founder and chief operating officer. This summarizes a more profound and visible industrywide shift in agentic AI.
Connectors with popular apps are proving important. Canva, for example, can link to Google Workspace apps as well as Slack, ChatGPT, Cloud, and Notion. Adobe also has some connectors, including ChatGPT, Dropbox, and Slack. Adobe’s Greenfield says AI agents “work like capable teammates”.
The versatile “small” models are also less expensive to use than the Frontier models, making it viable for enterprises to run agents on longer or continuous tasks. They can be trained for specific tasks. Obrecht points out that his models are five to thirty times cheaper than competing Frontier models. “This allows us to provide power to a large portion of our site, even for free users,” he says.
Investments in the AI Frontier Lab resulted in the cost-effective Canva Proteus, Canva Lucid Origin, and Canva I2V models. Overall economy means that agents work locally on tablets or laptops while keeping offline usage and privacy aspects in mind.
Anthropic’s Cloud Managed Agents, released this month, allows the platform to create autonomous agents that “can work autonomously for hours with progress and outputs that persist even after disconnection.” This can make agents relevant to complex workflows, which take time, accuracy and often require rework.
Google is investing in a core infrastructure called the Agent2Agent (A2A) protocol. A2A allows Google’s Gemini agents to “talk” to agents on other platforms. In a creative context, a team’s Adobe Firefly agent could theoretically send assets created according to specific guidelines to a Gemini agent for inclusion in a corporate presentation.