Better in English?
Better in English? Switch the website language for a better experience.
On June 24th, Figma published a sprawling changelog summarizing everything unveiled at Config 2026. And this time, it's not just another comfort update. It's a rethink of what you can actually do directly on the canvas.
Three major workstreams stand out:
In concrete terms, here are the updates that strike us as the most significant for an agency like ours.
Until now, prototyping a smooth micro-interaction often meant stepping outside Figma entirely, heading to After Effects, Principle or Framer, and then trying to translate that back into CSS or JS on the dev side. Figma Motion changes the game: you can now build precise, expressive animations directly in a timeline, or let the Figma agent generate them from a text prompt.
The big idea here: build reusable animation systems (the same way you'd build a component design system) and hand developers animations that are ready to ship.
3D transforms let you add depth to your designs and images, and combine them with Figma Motion for a result that moves like it would in real life. Shaders, meanwhile, work through description: you tell the agent what you want, and it generates the texture, ready to share and drop straight into a file.
Both features are currently in invite-only beta.
This is arguably the biggest shift of the bunch. The Figma agent can now be given "skills" (added capabilities), attachments and connectors that give it more context to work the way you do. It's also making its way into FigJam and Slides, where it can generate tables and diagrams, or edit decks in bulk.
And the cherry on top: generative plugins. You describe the tool you need, and the agent builds it for you, on the fly.
Let's start with the good news. Figma Motion is a genuinely great addition. At Mazette.co, we regularly hand off fine-tuned interactions to our Webflow developers, and the gap between "what we imagined in Figma" and "what's actually technically feasible" has always been a source of friction. Being able to prototype an animation right in the canvas, with a reusable system and dev-ready export, closes a real pain point. We see it as a concrete boost to our Figma-to-Webflow workflow.
3D transforms and Shaders look appealing on paper, but we're staying cautious here. These are beta features, heavy on creative flair but not necessarily essential for 90% of the projects we run. We'll keep a close eye on them without rushing in.
Where we're most torn is on the customization of the Figma agent. Giving it "skills" and connectors so it better understands a project's context is a smart move. But let's be honest: the more autonomous and connected to our internal tools the agent becomes, the more questions arise around governance and quality control. An agent that generates mockups or slides from your documents is a real time-saver, provided a human still keeps an eye on it. At Mazette.co, we treat AI as an accelerator, never as an autopilot.
Our overall verdict: this wave of updates confirms a trend we've been tracking for months, Figma is aiming to become the single platform where design, motion, code and AI all coexist. For an agency like ours, working across both design AND Webflow implementation, that's good news overall: fewer back-and-forths between tools, more context preserved along the way. It remains to be seen how these beta features hold up once they roll out widely, before we fully fold them into our migration and production processes.
Code
Over 120 companies have already stepped out of the shadows with us.
