Figma Bets AI Will Move Software Value From Code to Design
Figma co-founder and CEO Dylan Field told Bloomberg Technology that the company’s Config overhaul is meant to turn Figma’s canvas from a place for design artifacts into a broader production environment for software, motion, AI workflows and brand systems. Field’s argument is that AI will make code easier to generate, shifting value toward design judgment and human point of view rather than making designers less important.
Figma is trying to turn the design canvas into a production system
Figma’s platform overhaul is aimed at moving the product beyond a workspace for design artifacts and toward a place where teams build more of the digital product stack. Dylan Field described the change as an expansion of what can exist on the canvas: not only vectors and images, but interactive code, AI-generated workflows, motion, and advanced graphics alongside conventional design representations.
The practical shift begins with “code layers.” Field said Figma is adding layers that represent code running interactively on the canvas. A designer can take a design, move it into a code layer, and have it automatically converted. From there, the result can become interactive. Field pushed the idea further: in the future, he said, users could imagine prototyping or even going to production with code in Figma.
You can also in the future, imagine a world where you're prototyping or even going to production with code in Figma.
That distinction matters. Ed Ludlow introduced the overhaul as an “intelligent canvas” for full-stack digital creation that lets teams build animations, integrate AI agents, and write native code in one place. Field’s more specific account was that Figma is adding code representations on the canvas now, with production use framed as a future possibility.
The product also adds agent and plugin capabilities. Field said Figma agent plugins will let teams generate reusable workflows and run “skills.” A Figma interface shown during the discussion used the prompt: “Create a plugin that generates a multi-series area chart using my design system.” The example framed agents as tools for repeatable team workflows rather than only one-off generation.
Figma Motion is another part of the platform expansion. The demonstrated motion tooling included controls labeled “Linear,” “Curve,” and “Spring,” while Field said the company’s community was “over the moon” about the release. Asked whether designers had requested motion capabilities, Field said they had been asking for it since Figma started.
The overhaul also includes generative effects and shaders. Field characterized shaders as a way to do advanced graphics work directly on the canvas. The product set Ludlow described — animation, AI agents, and code in one place — is broader than a design-file upgrade. Field’s explanation positioned the canvas as the place where those production elements can sit next to the design itself.
The AI argument is that humans raise the ceiling
Ed Ludlow pressed Field on what had changed technologically to make the platform overhaul possible. Dylan Field moved from product mechanics to a broader argument about AI’s limits in creative work: models trained on in-distribution data tend to generate outputs near the average.
Field said that when users prompt AI systems, “so often” they get an average output, and “people can smell that.” He compared it to AI-written prose, where readers sometimes react by wondering whether a passage was generated. Ludlow agreed, with the qualification that the feeling depends on the platform and that not all systems are equal. Field said the same dynamic applies to design.
Only humans can really raise the ceiling here. They can use AI to be more effective. But humans bring the point of view.
The consequence, in Field’s telling, is that businesses will need stronger points of view, not less. As AI floods the attention economy with more content, design becomes more important as a differentiator. Companies trying to stand out will need to be bold across software, brand, marketing, advertising, and product experience. In his formulation, AI can make people more effective, but the value of the tooling is that it lets designers apply human judgment across more of the production stack.
Model modularity supports a larger claim about where value moves
Asked whether the coding layer was built by Figma itself or whether the company is model- or coding-platform-agnostic, Dylan Field said Figma has “always taken a modular approach” to models. The company can swap models in and out on behalf of users or at a user’s request.
That modularity matters because Field sees code itself becoming less scarce. His claim was not merely that Figma can call different AI models; it was that as code “commoditizes,” value moves up the stack to design. In that context, the code layer is also a claim about where software creation is headed. If code generation becomes easier and more abundant, Field argued, the differentiating work shifts toward design.
Ed Ludlow framed one business implication bluntly: “Tokens equal revenues.” Field’s answer was that Figma is excited to be “part of that token flow” because of what it can mean for users and their ability to affect the platform.
The more expansive version of the claim is that designers may move further into software creation. Field said Figma believes that in the future “designers are the ones” who will be building all the software. The statement depends on the earlier premise: if AI makes implementation more accessible, the person with the product judgment and design intent can move further into creation instead of handing off a static design to a separate implementation process.
Brand assets and workflows become part of the system of record
Ed Ludlow brought in an audience question about whether Figma plans to become a platform for brand integration using MCP. The question described a user who might build a website in another application but still wants an AI agent to access a “ground truth”: owned IP, branding, logos, artwork, and related assets.
Dylan Field answered more broadly than brand integration alone. He said Figma wants to do more for digital asset production workflows, including systems for advertising and media generation. The company’s bet here is Figma Weave, which Field described as a platform for composing models and workflows while preserving human craft.
The metaphor he used was shaping model output “like clay.” Weave, in his description, is meant to let users compose workflows, keep a human touch, and reach higher-quality outputs than a raw prompt might deliver. Those workflows can then be run through an API and pulled into other systems.
That answer positioned Figma not just as a repository for visual assets, but as a system of record for how those assets and brand expressions are produced. Field said the company is excited to be the system of record for brand assets, design systems, and the workflows themselves.
The distinction is important. A brand system in this framing is not only logos, colors, and reusable components. It also includes the repeatable processes that generate ads, media, product experiences, and other digital outputs. If AI agents are going to act on behalf of teams, Field’s answer suggests Figma wants them to draw from a system of record for brand assets, design systems, and workflows.