Canva unveils Canva AI 2.0, recasting its platform as an agentic system for work

Visual communication platform provider Canva Pty Ltd. today unveiled Canva AI 2.0, an overhaul of its platform that offers a conversational, agentic system aimed at becoming the place where teams start and end their workday.

The release, announced at the company’s Canva Create event in Los Angeles, is being pitched as Canva’s biggest shift since its 2013 launch, when it moved design from desktop software into the browser.

Canva AI 2.0 has been designed to transform Canva into a conversational, agentic platform where teams can go from idea to execution in one place and to deliver a creative partner across the entire process, from the spark of an idea to the final output.

The new offering is built around four capabilities.

The first one, Conversational Design, generates fully editable designs from natural language prompts, while the second, Agentic Orchestration, coordinates Canva’s suite of tools to execute multistep goals such as building a multichannel campaign from a single brief.

The third capability, Object-Based Intelligence, allows for precise edits to individual elements without regenerating the full design. The fourth, Living Memory, provides persistent context that learns a user’s style and brand preferences over time.

Alongside the core platform advancements, Canva also introduced a set of workflow features designed to pull outside work into the editor. They include connectors that link Canva to Slack, Notion, Zoom, Gmail, Google Drive and Google Calendar to allow the assistant to draw on emails, transcripts and calendar entries to generate briefing documents, sales pitches and newsletters. A new scheduling feature runs tasks in the background on a set cadence and web research pulls structured, editable information into designs on demand.

Other additions include Brand Intelligence for automatic brand application, Sheets AI for structured spreadsheet generation and Canva Code 2.0, which now supports HTML importing so users can bring AI-generated code into the visual editor without rebuilding it.

Powering the new version of Canva AI is the company’s in-house research group, Canva Original Research and Exploration, which the company said now employs more than 100 researchers.

Canva claims its proprietary models are up to seven times faster and 30 times cheaper than comparable frontier alternatives. The models include Canva Proteus for style transfer, Canva Lucid Origin for image generation and Canva I2V for image-to-video.

Lastly, Canva now allows artifacts generated by outside AI tools to be imported directly into its editor, where they become editable and collaborative through the platform’s drag-and-drop interface and can be published as functional websites. The move is in response to a growing pattern of AI platforms integrating with Canva to give users a path from raw generations to structured, scalable output.

Canva AI 2.0 is launching as a research preview today, before initially rolling out to the first 1 million users who discover it on the Canva homepage, with broader availability in the coming weeks.

Image: Canva