Naming a category, not just a product.

The first version of the landing page was a competent SaaS template. Feature pillars, persona tabs, comparison table, pricing cards, FAQ accordion. Every Figma plugin, every AI tool, every indie SaaS ships that exact skeleton. For designers with high visual literacy who scroll past dozens of these weekly, it registers as “another one” before they even read the copy.

The irony was sharp: a tool that authors bespoke design systems was presenting itself through the most generic possible design system.

The solution was a self-authoring page. The visitor opens to an almost-empty canvas: just the ] mark and a blinking cursor in a chat input. A prompt types out: “Build a landing page for a design tool. Warm, craft-forward, monospace accents.” Then the page builds itself. CSS custom properties cascade in and the Atelier palette transitions from nothing to the full brand identity. Sections materialize as if the AI is placing them. Typography locks in. Components snap to grid.

On intentstudio.design, the page you land on is the output. The medium and the message are the same thing.

Three layers run simultaneously. A chat simulation types the prompt with realistic timing. A variable cascade panel shows tokens appearing in real time (--color-primary: #B8741A) while the page's actual styles update in response. And sections materialize with a specific choreography: frame appears empty, auto-layout snaps children into position, then content populates. The sequencing is what makes designers lean forward. Too fast and it feels preloaded. Too slow and people bounce.

The landing page demonstrates the output. The Figma plugin is the engine. Intent Studio is a conversational AI tool that writes production design systems directly to the Figma canvas: variables, token hierarchies, components with proper bindings, documentation. Not code-first prototypes. Not wireframes. Native Figma elements authored through natural language.

Nothing else in the market does this. Figma Make generates code-first layouts, 3 frames max. Relume builds sitemaps externally. Every other AI tool in the space skips Figma entirely or treats it as an export target. Intent Studio treats Figma as the authoring environment, because that's where designers actually work.

The brand system has two halves that serve completely different purposes but reinforce each other everywhere they appear. The mark is a symbol. The voice is a language. Together they form an identity that speaks natively to the audience's daily vocabulary.

Seven colors. Three typefaces. No dark mode. The palette is named Atelier because it is meant to feel like a workshop, not a dashboard.

Before designing anything, I audited the visual identity of every tool in the space. The finding was striking. Every single competitor lives in cold blues, purples, and neutral grays. The entire warm end of the spectrum was completely unoccupied.

The product started as “Schemely.” It sounded like a wireframing tool. Every demo required me to first correct the assumption that this was some kind of layout generator. The name was creating an expectation gap that marketing would have to constantly fight against.

Building a production design system in Figma takes 8 to 12 weeks with a two-person team. Getting it into code is closer to six months. For freelancers juggling multiple clients and small agencies running lean, that timeline is the difference between delivering a proper system and delivering a pile of unconnected components.

I mapped every tool in the AI + Figma space and found a clear, open niche. Figma Make generates code-first prototypes, 3 frames max. Relume builds sitemaps externally. Figr Identity uses click-based UIs. Lovable and Bolt skip Figma entirely. Nobody was building conversational AI that writes native design system elements directly to the canvas.

The market data backed it up. 30% of Figma's highest-value enterprise customers were already using AI-assisted design weekly. 61% of small-company Figma users said AI was critical to staying competitive. The demand existed. The tool didn't. And based on Figma's code-first trajectory, the window to establish this category was 12 to 18 months.

The target audience is freelance designers and small agency teams. People who already know what design systems are. They don't need the concept explained. They need to see that this tool understands their workflow and respects their time. The voice principles follow from that.

The go-to-market is built around one bet: designers try tools when they see the output, not when they read the pitch. Every channel exists to get the demo in front of the right people. Everything else is distribution for that moment.

The model is land and expand. Land with a free tier that lets individual designers author a single component set on real work inside five minutes. No credit card. No onboarding flow. Just open the plugin and describe what you want. Expand into paid when they hit the second or third file and the monthly credit allotment becomes the obvious next step.

Channel strategy is three lanes. A steady drip of short-form case studies on LinkedIn aimed at agency design leads. A small set of long-form writeups on the product blog for the IC product designer audience who wants to understand how the token architecture works.

The rename was the single highest-leverage decision in the entire project. “Schemely” required every conversation to start with what the product isn't. “Intent Studio” lets the conversation start with what it does. I underestimated how much a name that creates the wrong expectation costs in ongoing correction. No amount of marketing spend recovers that.

The visual audit changed how I think about positioning. I went in expecting to find a color I liked and came out understanding that color territory is a strategic asset. Every competitor in this space uses some variation of blue, purple, or neutral gray. Moving to the warm end of the spectrum wasn't a preference. It was a defensible position. You can see Intent Studio in a plugin panel or a social feed and know it's different before you read a word.

The --intent: voice proved a notion I believe strongly from my years of agency work: the best brand languages aren't invented from scratch. They're borrowed from the audience's own vocabulary and given a new context. CSS custom properties are something designers and developers write every day. Turning that syntax into a brand expression made the whole identity feel native instead of imposed. It wasn't clever for the sake of being clever. It was just the right language for the audience.

The CSS-native brand voice achieved immediate resonance with technical audiences in a way that a conventional tagline never would have. When your brand literally speaks the same language your users type into their editors, the gap between “marketing” and “product” collapses. That was the goal.