Words Are the Design

Recently, I was walking through a product journey with a prospective client. They wanted to optimize their onboarding workflows with AI agents, qualify the signup, route each user to the right path, and follow up automatically. Straightforward work. I have the templates and the skills in my Design AI toolbox for exactly this; we could have had agents running inside a week.

That's not what gave me pause. The copy did. The welcome screen didn't sound like the brand. The empty states read like placeholders someone forgot to replace. Three screens described the same feature three different ways. The words felt last-minute, written after the design was "done," by whoever had ten minutes and a template. Much as I hate to admit it, I've been up late at night, blurry-eyed, writing error copy at the last minute while an engineer waited to go live.

Here's the problem with optimizing that flow: AI agents don't fix unkleer langwidgeunclear language. They distribute it. Automate that onboarding, and you aren't delivering a better experience faster; you're delivering confusion at scale. The workflow was never the hard part. The words were.

How much of your product is just words?

Strip them out and look at what's left: a collection of colored rectangles that don't mean much. Labels, headers, CTAs, error messages, empty states, onboarding flows, confirmation dialogs — most of what a product actually is, is text. (Great products are text and a strong POV.)

And yet design culture, at some point, decided that "real" design was primarily visual. The layout gets three rounds of review. But the button label gets picked in a Slack message on a Friday afternoon.

The first things I look at are copy, font hierarchy, whitespace, and color; my feedback is almost never about the radius of a corner. It's about meaning. Does this screen say what it's for? Does the user know what happens next? Words first, then everything else.

That hierarchy was never a personal quirk. The history of advertising has been trying to tell us this for a hundred years. The greatest agencies in advertising history weren't founded by art directors. They were founded by people who could explain things clearly and make you feel something. That combination built empires. It still does. We just call it something different now.

Look at who actually built the most enduring agencies:

Ogilvy & Mather — David Ogilvy, copywriter. The self-described "Father of Advertising." Wrote the Rolls-Royce headline. Built one of the most awarded agencies in history.

Leo Burnett — Leo Burnett, started as a reporter. Created the Marlboro Man, Tony the Tiger, the Jolly Green Giant. Built the agency that grew into a global network of thousands across dozens of offices.

Abbott Mead Vickers — David Abbott, copywriter. Started at Kodak. Co-founded what became the UK's largest agency.

Wieden+Kennedy — Dan Wieden, copywriter. Wrote "Just Do It." His father was a copywriter. Copywriting wasn't just his background — it was his inheritance.

These weren't people who wrote taglines on the side while the real designers worked. They ran the shop. And the common thread wasn't talent, exactly; it was a specific kind of capability: the ability to explain something so clearly, and make you feel something so precisely, that people couldn't look away.

They sold products with words. Product designers build products out of words. The skill is the same. The stakes are higher.

I came up on the words side of this myself. My first company, EatAgency, was a content strategy shop. We wrote for Walden University, McKinsey, Laureate Education, and SAP. The work pulled us into UX and product, building for American Express, Kindling, JinglePunks, and DAV. Nobody handed us a new skill set when we crossed over. The clients changed. The deliverables changed. The job was always the same: explain this clearly, make someone feel something, get them to click a CTA.

And yes, those famous agencies didn't build products. They made campaigns. A beautiful headline and a finished app are different things. But at the core it's the same thing. As product designers, we design for emotion, anxiety mostly. A person has a problem, some amount of pressure to solve it, and a range of understanding on how. Product designers manage that equation, and words are how you manage it.

Ogilvy wrote copy about products. Product designers write copy inside products. An ad is separate from the thing it's selling. But an error message, a navigation label, an empty state: those are the product. There's no distance between the word and the experience. When the language fails, the experience fails. When the language works, the product becomes legible, learnable, trustworthy.

A campaign is a representation of a product. An interface is the product. Words in advertising sell. Words in product design instruct, orient, reassure, and, when they're working at their best, disappear entirely so the person can just do.

A campaign is a representation of a product. An interface is the product.

The stakes are different. The underlying skill is the same: precise articulation in service of an outcome.

The skill that matters most in an AI-native workflow is prompting. And prompting is very disciplined writing. It's knowing how to frame intent. How to give context without noise. How to ask for something specific enough to get it, and open enough to be surprised by what comes back.

The people extracting the most from AI right now aren't always the most technical. They're the most articulate. They write prompts the way Ogilvy wrote briefs: with a clear point of view, a defined audience, and a desired feeling at the end. They know that clarity of input determines quality of output. They treat imprecision as a bug to fix, not a style choice.

Ogilvy would have been exceptional at this. Wieden too. Not because they were tech-savvy, but because they spent their careers thinking about how to communicate intent precisely and emotionally.

If you can explain something well, you can design it well. You can brief it well. You can prompt it well.

This isn't hypothetical. In the last twelve months, most of what I've been brought in to fix has been a messaging challenge, a degradation of the promise and the product. Teams shipping faster than ever, and the words falling apart first. The product stops saying what it's for. Trace those failures back, and they land in the same place: copy. Words.

This isn't a call to make designers into writers. It's a call to stop treating language as a downstream handoff and start treating it as a core design material at the same level as color, spacing, and hierarchy. The designers who rise in an AI-accelerated environment won't necessarily be the ones with the deepest Figma fluency. They'll be the ones who can articulate a problem so precisely that it becomes solvable, who understand that every label, every instruction, every empty state is a design decision with a user on the other end of it.