Design as Infrastructure

Dear Teams that ship,

AI coding assistants change the fundamental economics of Designers on production teams. It enables designers to deliver functional component libraries and design systems in code—not mockups that need translation, but working components that ship to production. How will this design as a role?

The Translation Tax

Most design teams still work the old way: create mockups, export them, write specs, and hope developers implement correctly. Every handoff creates friction. Mockups become specs. Specs become tickets. Tickets become implementations that rarely match the original vision. Each step introduces delay, miscommunication, and compromise. This wasn't anyone's fault—it was the most efficient approach given the capabilities.

Organizations treated design as overhead because design created overhead—every deliverable required translation.

That translation layer is now optional.

The Opportunity

The design engineer role isn't new. The industry attempted this convergence repeatedly—but failed to achieve mass adoption because the cognitive load was unsustainable. Being excellent at visual design, interaction design, and production-quality frontend, required expertise across domains that few could maintain simultaneously. The separation was economically rational given the constraints.

AI coding assistants have dramatically changed this equation.

Here’s a simple example that highlights what’s possible now: A designer describes a navigation component to an AI agent—sticky header with precise dimensions, responsive breakpoints, smooth shadow on scroll, and keyboard navigation. The agent generates the React component, the designer reviews it, adjusts the easing curve, handles edge cases, and creates a PR.

No translation required, no fidelity loss.

With our AI Design Partners

Design exploration and discovery are expressed through highly iterative cycles in which designers use AI for research, testing, and enriching metadata and content for their designs, helping them better understand the real-world needs of the problems they’re solving.

In production, designers now ship what developers need: functional components with built-in interaction logic, accessibility features, responsive behavior, and edge cases. Not just documentation—but design tokens in code that globally enforce consistency.

The design system isn't kept in Figma as a reference anymore—it's shipped to production codebases as logic and rules for AI assistants.

Designers become systems architects who build and maintain foundational components and ensure the quality of the product experience through implementation.

When designers deliver functional components rather than requiring translation, design stops chasing QA sessions and documentation in the later stages of production.

Designers who embraced coding before AI arrived will lead this shift, working close to the code. Most importantly, this enables designers to craft the experience from conceptualization through to implementation.

The Change

The capability shift has already happened, entire companies are being reconfigured to embrace new AI opportunities, but many design teams are still organized around Figma deliverables and complex developer handoffs—optimizing for a constraint that no longer exists.

Our design programs and tooling will evolve to better meet the design infrastructure needs of our faster-moving, leaner, and more experimental product teams.

Designers have a generational opportunity to evolve their craft away from expensive translation layers and move toward deeper integration and ownership of the entire customer experience.

Long term, design teams that deliver and maintain functional systems in code, at scale, become infrastructure. The rest become optional.