Figma Conference: CEO Dylan Field Says AI Must Solve Real Problems, Eyes Design-to-Prototype Leap

Dylan Field, CEO of Figma (NYSE:FIG), said the company’s approach to artificial intelligence has centered on applying the technology to “solve real problems,” rather than adding AI as a marketing layer. Speaking in a conversation about how product design is evolving in the age of AI, Field pointed to early, narrow use cases—such as automatically renaming design layers—as examples of features that saved designers significant time and were immediately valued by users.

From small AI utilities to “design-to-prototype” workflows

Field said Figma’s broader AI ambitions are reflected in products like Figma Make, which he described as enabling users to move “right from design to prototype to back to design,” arguing that designing and prototyping “should be two sides of the same coin.” Looking ahead, he said Figma is focused on bringing “all the efficiency of experiences you can get in code to the designing canvas,” signaling interest in capabilities that narrow the gap between design and implementation.

He also described a future where designers could do more than create prototypes, potentially reaching a point where design work can help update an existing code base. Field cautioned that time frames are uncertain, noting “in 6 months, 12 months, it’s hard to say exact time frames because we don’t know.”

Roles may persist, but responsibilities will blur

Asked whether AI-driven tools will cause product roles—designer, product manager, engineer, research—to fuse together, Field said his view has shifted. He previously believed roles would blur, but said his updated take is that roles will remain while responsibilities blur. In his view, AI pushes teams to be more generalist and reduces the ability to stay blocked behind role boundaries.

Field said people will increasingly take on tasks outside their traditional lane:

  • More people will prototype and participate in design.
  • More people will engage with customers and “do all of it.”
  • Specialized expertise will still matter, which is why roles persist.

Agents, interfaces, and the importance of auditability

On the future of user interfaces in an agent-driven world, Field said current thinking about agents is “limited,” typically framed as a user launching a task and receiving a finished result. While he expects more autonomous, long-running execution over time, he also predicted the “prompt box” will evolve.

Field compared the current stage of AI interaction to an “MS-DOS era of AI,” suggesting the industry has not yet arrived at more mature interface paradigms. He emphasized that as agents become more capable, auditability will remain critical—people need to be able to review and understand what an agent did, rather than blindly trusting outputs.

He also said the future is likely less about agents working independently “in a sandbox” and more about agents and humans working alongside each other in collaborative environments. Pointing to Figma’s multiplayer canvas and inspiration from Google Docs-style collaboration, Field argued software will move toward a “multiplayer philosophy,” where humans and agents operate together, even if agents sometimes run tasks independently over long periods.

Judgment, point of view, and “opinionated” outputs

As AI enables near-instant generation of many design variations, Field said the key challenge shifts toward framing problems correctly and applying judgment. He argued that producing an output from a prompt is only a first step; prompting well is difficult, and teams still must ensure consistency across “so many surfaces” and modalities. Most importantly, he said design needs a point of view to stand out in an increasingly competitive software landscape.

Field added that “opinionated” outputs matter because they generate a stronger reaction and serve as better material for iteration. In contrast, he said generic “AI slop” may technically work but fails to provoke a meaningful response. He described outputs as “clay” that still needs shaping through a process.

In that context, Field referenced an acquisition, saying Figma is working with a company it acquired called Weavy, now called Figma Weave. He said the goal is to enable users to manipulate media outputs through a repeatable process that yields “outstanding” final assets.

Riding foundation model progress and measuring ROI

On whether Figma aims to build its own models, Field said the company experiments broadly, but emphasized that foundation models are improving so quickly it would be “foolish” to assume a company can outpace them. Instead, he said the strategic imperative is to “ride that wave” and add value on top of it.

When asked about ROI metrics for AI and Figma, Field said he expects the time from design to development to shrink significantly. However, he stressed that craft, design, and taste are “inherently non-verifiable aspects,” making them hard to quantify. He suggested organizations should focus on qualitative decision-making processes—specifically, who decides when a product ships and how the tradeoffs between speed, business constraints, and quality are managed.

Field argued that companies risk moving faster while shipping worse products, and said cultures that do not treat design as part of the “gate” for shipping may struggle more. He urged organizations to prioritize quality and ensure designers have a meaningful role in determining when work is ready to release.

In closing remarks, the interviewer noted plans to partner closely with Figma to “build great software,” citing a desire for products that “move people emotionally.” Field thanked the interviewer for the partnership and concluded the discussion.

About Figma (NYSE:FIG)

Figma is a San Francisco–based software company that offers a web-based platform for interface design, prototyping and collaboration. Its flagship product, Figma, enables teams to create and refine user interfaces, vector graphics and design systems directly in a browser, eliminating the need for local installations. The platform’s real-time collaboration features allow multiple stakeholders—designers, developers and product managers—to edit and comment simultaneously, streamlining workflows and reducing version control issues.

In addition to its core design tool, Figma provides FigJam, a digital whiteboarding solution that facilitates brainstorming sessions, wireframing and diagramming.

Recommended Stories