Canva's AI 2.0 brings prompt-powered design to the browser
Canva just overhauled its design suite with AI 2.0, turning the platform into a conversational, agentic workspace. Users can generate multi-channel campaigns or tweak specific design elements using plain text prompts, backed by persistent brand memory. The update is rolling out as a research preview to the first one million users on Canva's homepage.

Anyone who has tried designing with AI knows the exact wall you inevitably hit: you feed the tool a prompt, it spits out a slick image, and the moment you want to fix a single typo or nudge a photo two inches to the left, you realize the entire canvas is baked flat. You are stuck rolling the dice with another prompt.
Canva is done with that approach. The company just rolled out Canva AI 2.0, describing it as its biggest platform overhaul since moving desktop publishing into the browser.
The update is launching today as a research preview for the first one million users who land on the Canva homepage, with broader access expanding over the coming weeks.
An orchestration layer that keeps elements editable
The core of the upgrade is a new orchestration layer that connects Canva's disparate tools behind a single conversational prompt box. Instead of manually opening five different templates, you can tell the chatbot to build a full multi-channel launch campaign for a summer product line, and it will generate the complete batch across multiple formats.
Crucially, what comes out the other end is not a flat JPEG. Through a feature called Object-Based Intelligence, every generated asset remains broken down into native, editable layers.
You can use plain text commands to swap a specific font, tweak copy, or change an individual image without disturbing the surrounding layout. According to Canva co-founder and CPO Cameron Adams, these dynamic assets lean directly on technology from Canva's 2024 acquisition of Leonardo.ai.
To make this work, Canva trained its in-house Canva Design Model not on finished images, but on the step-by-step editing actions of its 265 million-plus monthly active users. The model understands the hesitation, the font swaps, and the layout pivots that happen between a blank canvas and a finished asset. It even outputs its creative reasoning alongside the design. AI 2.0 also introduces persistent memory to keep team brand kits consistent, supports HTML imports in Canva Code, and plugs directly into Slack, Gmail, Google Drive, and Google Calendar.
Why it matters
We think Canva made the exact right bet here: the battleground for AI design is no longer raw image generation, but post-generation control.
Generating a pretty picture in 2026 is table stakes. Being able to hand off an actual asset that a marketing coordinator can fine-tune in thirty seconds is where enterprise software lives or dies. If you run a small business or build decks on Canva daily, getting onto the one-million-user waitlist is an absolute no-brainer.
This update also lights a massive fire under Adobe, which announced its own prompt-based editing features just yesterday. While Adobe is building downward from complex pro software into prompt controls, Canva is building upward from simple browser templates into full creative agents.
The catch is whether Canva's reasoning can truly grasp subjective creative nuance. Telling an AI to create tension in negative space sounds incredible in an interview, but executing it without cluttering the canvas is a much harder engineering problem. Still, as pure execution becomes free, raw creative judgment is officially the only moat left.
Sources (1)
- theverge.com(opens in a new tab)Originaltheverge.com