This essay first ran on The Brand Architect, our Substack, on February 11, 2026. It is republished here, lightly edited. Since then, the “framework” it describes has a second product: Kitana Lite, our brand review agent.
In 1919, Walter Gropius opened the Bauhaus in Weimar with a simple thesis: industrialization had changed how things were made, and design needed a new framework to keep up. Factories could now mass-produce furniture, textiles and architecture, but without a philosophy for designing inside that capability, you got ugly mass production. The Bauhaus didn’t reject industry. It built the principles for designing within it.
AI is doing the same thing to content and design right now. Anyone can generate images, write copy and produce layouts at a speed that would have been unimaginable five years ago, and the frameworks haven’t caught up, which is why most of what AI produces looks like what it is: undirected output with no point of view.
Every major shift in technology creates this gap. The tools arrive first, and the philosophy for making meaning with them takes longer, because it’s usually builders and designers who work it out by using the material until they understand what it’s good for.
That’s what we’re doing at RepliHaus. We treat AI as a new material that designers can work with, shape and direct, rather than a replacement for design thinking.
The handoff nobody built for
I’ve spent my career in design: branding, typography, UI, 3D, product. I’ve been the designer people come to when something has to look right, and also the one who gets 14 Slack messages on a Monday morning asking for “quick” ad variations due by Wednesday.
If you run a design team, you know that feeling. If you’re in marketing, you know the other side of it: the waiting, the workarounds, and the “just use the template in Figma” ask that assumes you won’t break anything.
What’s strange is that the industry has spent years perfecting the handoff from design to engineering: Figma to code, design tokens, component libraries that sync with production. The handoff from design to marketing is almost ignored, and the tooling between the two is “send a Figma link and hope for the best.”
Template libraries exist, and designers are still bottlenecks, because none of these tools solve the trust problem. Designers don’t hand over the keys because they’ve seen what happens when they do: one stretched logo or one “that color looked close enough” and months of brand work unravel.
And now there’s a newer version of the same problem. Most companies already use AI for creative work, whether they admit it or not: someone in marketing runs prompts through ChatGPT, someone else generates images in Midjourney, and it all happens in the shadows, uncontrolled and off-brand. The volume of content a team can produce just went up by an order of magnitude, and the governance didn’t move at all.
This is the Bauhaus problem a century later. Production capacity has outpaced the design thinking.
What a framework for this era looks like
If the problem is a one-way handoff, the framework has to make the handoff between design and marketing go both ways.
Design gives marketing templates that non-designers can use without breaking anything. That means no design software to navigate, just a form: fill in the headline, pick the image, export. The rules are built in, so the margin for error is gone.
In return, marketing gives design something it has never had: data. Which templates work, which don’t, and why. Right now designers set rules and never hear what happened, and marketers request assets and never say what performed, so both sides work in the dark.
AI makes both halves possible for the first time: templates that enforce brand rules on their own, and performance feedback that flows back to the people making creative decisions. A loop instead of a one-way handoff.
Where RepliHaus starts
RepliKit is our first product and the first half of that loop. It’s a Figma plugin: a designer builds a template and tags the editable parts, and everything else is locked. A marketer then fills in content through a form without ever touching Figma, and the system enforces the designer’s rules on every export. The marketer gets their 20 ad variations by Friday with no design review, because the template made going off-brand impossible.
It’s in paid alpha, and it works. But it’s only the first expression of the philosophy.
From here we’re building toward the second half: templates that learn from how they’re used, AI that understands your brand because it trained on your designers’ work, and a platform that feeds the data back so both sides keep improving.
Why this matters now
Brands need far more content than they did a few years ago: more platforms, more formats, more markets, more personalization. The teams making that content haven’t grown at the same rate, and AI is about to push the volume up again, which makes brand governance more urgent just as it gets harder.
The Bauhaus answered industrialization by saying: we don’t reject the factory, we give it a design philosophy.
That’s our bet. The companies that learn to design with AI, inside its constraints and with real intention, will produce work that means something. Everyone else will drown in generic output.
We’re building the framework for the first kind.
If this resonates
If you’re a design leader tired of being the bottleneck, or a marketing leader tired of waiting in the queue, I’d like to hear from you. I don’t mean a sales pitch, just a conversation about how your team works and where it breaks down.
We’re at replikit.ai, and I’m at hey@seun.design.
First published on replikit.substack.com. Topics: design, AI, brand systems.
