Brief IA

Figma and Agentic AI: A Revolution for Designers in 2026

🛠️ AI Tools·Tom Levy·

Figma and Agentic AI: A Revolution for Designers in 2026

Figma and Agentic AI: A Revolution for Designers in 2026
Key Takeaways
1Brad Frost and Dominic Nguyen demonstrated the potential of agentic design systems during a live session.
2AI can now create components from existing bases, providing unprecedented creative control to designers.
3Fundamentals like semantic tokens and consistent nomenclature become crucial for effective collaboration with AI.
💡Why it mattersThis evolution redefines the role of designers, giving them more power and efficiency in the creative process.
Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisis ton rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

📄
Full Analysis

A Revelatory Demonstration

The foundations of Figma that you mastered are now the bedrock of a major transformation. During a striking demonstration, Brad Frost, author of the influential book Atomic Design, teamed up with Dominic Nguyen, co-founder of Chromatic, for a live session titled Agentic Design Systems in 2026. This session was hosted by Kyle, a developer experience engineer from the Storybook team, who presented their ongoing MCP development.

As a designer and educator, I often participate in such demonstrations to stay informed, although I often see myself as an observer in others' fields. However, this demonstration captured my full attention. Kyle executed a simple command: “Add a customer review component.” At that moment, no review component existed. There was no design file, no ticket, no Figma frame to reference. The agent identified a Star component, Typography, and an Avatar. It analyzed their properties, understood their states, and composed a new element, wrote the code, and drafted the necessary tests. All of this took less time than it would have to write a Jira ticket. This demonstration was primarily aimed at developers, but as a designer, my reaction was quite different.

A New Era for Design Systems

Design systems are no longer just simple documents for developers. They are transforming into true instructions for machines. The person who decides which components exist, what states they have, how they are named, and why, is the designer. This is not a threat to the profession, but rather an opportunity to exercise unprecedented creative control over what is built. What Storybook illustrated was not an AI creating design, but an AI that finally knows how to utilize what the designer has already crafted. This is the opposite of vibe coding. The agent does not create new things; it follows established guidelines.

The Fundamentals Not to Overlook

It is crucial to note that nothing making this advancement possible is truly new. Semantic tokens, a consistent nomenclature, and complete component states are concepts that design system experts have advocated for years. Most of us have politely listened to this advice before returning to our daily work, often because it seemed to be advice meant for large companies and over-engineering for a small under-resourced team trying to deliver a project by the end of the week.

The difference today lies in the fact that the end user is sometimes a machine that reads exactly what is written. The foundations of the design system that we have always been advised to master in Figma are now the basis of an agentic configuration where designers could have more power than ever over the creative outcome, provided it is well-directed.

Good News: Figma is Not Dead

For those wondering, the site isfigmadead.com confirms that Figma is still very much alive. What truly changes in an agentic configuration is the structure and where you focus your energy. It’s about spending less time on full-page mockups, often viewed only once during a handoff meeting, and more time on the building blocks: components, tokens, states. This represents a true return to Brad Frost's atoms, who has likely been pleased to see his work become the infrastructure of an AI workflow.

Deliverables and the Process

A) The Deliverables

The agent builds from the building blocks: components, tokens, styles, and the descriptions you write for each. These are not new practices, but the good old fundamentals of the design system, executed correctly.

B) The Process Towards Solid Deliverables

Good building blocks do not come from a component library maintained in isolation. They come from a designer who has first considered the whole. You cannot extract a good Star component without having designed a page where the star was meant to function, where you saw how it related to the surrounding typography, how it integrated into the context, and what it communicated.

The design system is a distilled discovery. Designing the entire page is where the discovery took place and is therefore not redundant.

File Setup

Now that we have defended the creative soul of design, let’s talk a bit about naming conventions. Think of this as a checklist for your file.

  • Your variables need more than just a color palette.
  • Match your component properties to the props in the code.

It is essential to establish a conversation with your developer before building anything. Property names and values must be agreed upon by both sides.

Conclusion

The creative process is not an overload. It is messy, non-linear, and impossible to automate. The depth of this mess is exactly what makes the distillation we feed to the machine valuable. If you cut it, you have a fast system producing forgettable results, which costs more in the long run than the time you thought you were saving.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.