More designers are now building for AI than ever before. According to Figma’s latest report, over half of users are working on AI products—up from just 21% last year. That shift matters. Because it means agents—not just humans—are starting to use our design systems.

The catch? Most systems today are still built for people. Not machines.

And certainly not for agents trying to interpret inconsistent naming, vague hierarchy, or undocumented conventions. If your components are messy, you’re not just slowing people down—you’re building walls between your system and the future. That’s where naming comes in.

Naming is the new interface.

Clean, intentional naming helps agents understand structure, purpose, and how elements connect. In that way, your layer names start to act like an API—defining inputs, outputs, and behavior. This isn’t about “making it pretty.” It’s about making it usable by something that can’t see.

A few ways to build for that future:

• Use semantic names, not stylistic ones.

• Stay consistent—avoid guesswork.

• Make relationships explicit.

• Think like you’re designing for a parser, not a person.

We don’t need to reinvent our systems. We just need to refactor them for clarity. The agents will handle the rest.