By Francesco Bertelli

The Designer as Conductor in the Age of AI

A few days ago, I built a fully working web application — without writing a single line of code. I did it entirely through typed conversations with an AI assistant. (I’ll share the app itself in a follow-up post.)

What this revealed — something I’ve always sensed but never fully articulated — is that the tools we’ve idolized in design are often barriers, not bridges. Richard Avedon once said: “I hate cameras. They interfere. I wish I could work with my eyes alone.” He wasn’t being dramatic — he was articulating the dream: pure translation from vision to creation, without mediation.

For the first time, that dream feels real.

The Real Bottleneck

As a designer who codes, I’ve always lived at the intersection of visual thinking and implementation. I enjoy that space. But when AI coding first arrived, I dismissed it as hype. What changed my mind was an article about “vibe coding” by Andrew Chen — a practice where developers guide AI through natural language instead of writing code.

It reminded me of the early 2000s when APIs let us remix the web, or when PhoneGap in 2008 let me build iOS apps using plain JavaScript. There was a thrill to those moments: technical friction was dropping, and creativity was rushing in to fill the space.

That thrill is back. Only this time, the friction isn’t technical — it’s cognitive.

When Execution Melts Away

Within minutes of starting, I had a working prototype. User auth, databases, responsive UI, live API calls — the works. What should’ve taken weeks came together in an afternoon. It felt less like building and more like conjuring.

But this wasn’t magic. It was a new kind of work: thinking clearly, communicating precisely, and debugging through language.

One surprising shift: the cost of experimentation has dropped to near-zero. You can try five directions, discard four, refine one — all within a few prompts. There’s no build time, no Figma layers to organize, no codebase to wrangle. Just ideas, tested and reshaped in real time.

You’re not iterating on pixels. You’re iterating on thought.

From Craft to Clarity

The constraint has shifted. The question is no longer “Can I build this?” but “Can I articulate what I want with clarity?” That’s uncomfortable, because it exposes something many designers have hidden behind technical polish: fuzziness of thought.

We’ve spent years perfecting button styles, spacing systems, component libraries. That kind of digital craftsmanship had its place. But in an AI-driven workflow, those skills matter less than your ability to define intent, make decisions, and shape direction.

We’re not losing design. We’re stripping it down to its essence: thinking.

The Designer as Conductor

This isn’t about replacing designers. It’s about repositioning them.

The future designer won’t be the one who can code every micro-animation by hand. They’ll be the one who can see the big picture, communicate it crisply, and orchestrate a system toward that vision. They’ll be less like a craftsman and more like a director or editor. The person who knows what belongs — and what doesn’t.

Taste, judgment, clarity — these become the new essential skills. And frankly, they always were. We’re just now being forced to confront that.

Embrace the Friction Shift

To designers clinging to the familiar: this isn’t about losing control. It’s about a different kind of control. One that requires letting go of your comfort with tools and embracing discomfort with ambiguity, with language, with your own unformed ideas.

Try it. Start with something small. Build it conversationally. Let yourself feel the frustration, the disorientation — and then the exhilaration.

What you’ll discover is not just a new toolset. It’s a new mindset.

What Remains

Design has been democratized. Anyone can produce something visually competent with AI. The new differentiator isn’t execution — it’s vision.

When tools disappear, thinking emerges. When craft becomes cheap, taste becomes rare. And when execution is instant, the only thing that matters is the clarity of your mind.

We are, finally, designing at the speed of thought.

P.S.

There’s a growing ecosystem of AI-first editors and coding companions — tools like Cursor, Bolt, Codeium, and Cline are all exploring different angles on collaborative, AI-driven development.
My AI-built app cost ~$200 in tokens — most of it spent debugging a stubborn password reset flow.
I tested voice-based “vibe coding” using Superwhisper, a tool for speaking directly to AI. It’s fast and intuitive — but for now, typing still wins for clarity (even with typos).
This article was co-written with Claude Sonnet 3.7 and copy-edited by GPT-4.5-preview.

About the author

Francesco Bertelli is Group Design Director at Work &Co, with over 15 years of experience designing and developing award-winning work for clients including Syfy, Target, and American Express.

This article originally appeared on Medium.

500 Unexpected error

404