Meeting people where they work

Part 3 of Building Arbor, a series about my year on Linktree’s design system.

A design system only pays off if the right thing is also the easy thing, at the exact moment someone’s making the decision. Research that lives in a doc nobody opens mid-task doesn’t change behavior. So the piece that mattered most after the researcher was a Figma plugin that pulled the system’s intelligence directly onto the canvas, where designers already are.

It did two jobs.

Bulk token migration

Behind the scenes, I’d remapped Arbor’s tokens in Figma to point at our canonical values and naming conventions: the unglamorous cleanup everything else depends on. But a cleaner source of truth is worthless if migrating to it lands on every designer’s plate as a chore.

So the plugin could find every place a file used an off-system or outdated token value and move them onto the canonical tokens in a single sweep. Designers got the migration for free.

Drift is the thing every design system quietly loses to. This closed the loop between what the tokens say and what the file actually uses.

UX writing, in place

I embedded our UX-writing skill right into the plugin, so you could select a piece of copy on the canvas and reformat it through the house voice without leaving Figma. The guidance stopped being a document you were supposed to remember and became a button you pressed.

That’s a small change with a large effect. Having a voice guide and actually applying it are far apart, and most guidance never makes that trip. A single click is how a standard becomes the default.

The through-line

The instinct here is the same as with the researcher: take the system’s knowledge and deliver it at the point of use, in the tool the person is already holding, instead of asking them to go find it.

Which raised a question. I now had a research corpus, a set of design tokens, and component specs. Three sources of truth, three places you had to remember to check. The next post is about what tied them into one.