Canva Introduces Canva AI 20 Brings Agentic Capabilities and Memory to Perform Design Tasks

On Thursday Canva announced a major upgrade to its artificial intelligence (AI) suite, which can be introduced by the company’s founder, Canvas. Canva AI 2 – Dubbed canva artificial intelligence 2 by . In the new suite, a agentic system that understands layered design is added to 0, the latest suite provides an advanced solution for complex tasks through ‘chat-based interface of complexity’. This event, hosted at Hollywood Park in Los Angeles, California’s Canva Create 2026, was the first previewed suite for the suite. In the suite, there are new features such as multi-channel campaign design (multi-modal memory), web research and task scheduling; an overall experience that is improved.

Canva AI 2.0 Is Here

The Sydney-based visual communications platform, Canva AI 2 was introduced in a newsroom post on the website’s behalf by its Sydney counterpart. 0. All existing features are redesigned and improved in the upgraded AI suite, which includes new functionality and improvements across all of its components. It is still not yet available to subscribed users but it will be available globally in the next weeks, according to the company.

Canva AI Discussion

• Should I get a subscription to Canva? The AI features look interesting, but I’m not sure.

Explore More…

The greatest change is introducing an agentic layer across the entire suite of s. It also makes the Canva AI chatbot, from a simple generation tool to ‘an assistant of conversation that can do everything in one platform’. A person who has a new design goal can submit text descriptions, voice messages or share ‘an rough sketch or – i.e. an brief on the subject of vs now users and the agentic chatbot finds the right tools to bring the idea to reality with it.” In a number of steps, users can “iterate” on ‘generated design in multiple steps,” the company says that chatbot has now been shown to maintain context.

Canva’s AI chatbot also generates multi-channel design projects with one prompt, eliminating the need for separate prompts for different designs. It also has a new Magic Layers feature that allows the user to control if he created ‘an original design’ and edits across the various layers of each layer, which will produce the final output.

Canva AI can also edit other elements in a design, and its actions are limited to that single layer using the same architecture. This means that if the user wants chatbot to change text font or background lighting (only element changes), everything else is same, and all else remains the same. It is referred to as layered object intelligence, which means this ability.

The introduction of Memory Library is another new addition to the term “”. Essentially, the chatbot is persistent memory (as in most of its work), which stores user’s preferences, style guides and other information for future tasks. When I use the chatbot, Canva says that a user is more likely to be used for surfacing suggestions and creating designs.

Canva also is introducing six new connector-based workflows that allow the chatbot to use third party tools and their data to create designs and assets. This includes connectors for Slack, Gmail and Google Calendar (Notion & Zoom) and HubSpot.

It also has the ability to schedule posts in its AI suite, which can be added by Canva. In this way, users can set a time when they need the output of design (and even at fixed intervals) to repeat it. Even if the user is offline, chatbot still generates content.

It also will improve usability of AI suite with the new web research feature, which is already available in a browser. It is now possible for chatbots to collect and gather information from the Internet, which can be used in a structured, editable format of design projects by the user. In addition to this, the update also makes improvements to Canva AI’s brand intelligence, Canvas Code and Sheets AI.