How it works

Built by us, run by your team.

Brand OS runs on two loops. We set up, build and review the system that makes your work, and your team approves every asset.

Two loops, and a person in both.

We set up and build the system, then stay in the loop to review it and keep it sharp. Your team approves every asset.

The Brand OS loop Five steps in a row: brand system in Figma, brief, generate with Claude, review and ship. A dotted inner loop runs from review back to generate for your team on every asset. A gradient outer loop runs from review back to the brand system: Ótimo sets up the design system, then reviews the output and refines it every month. INNER LOOP · YOUR TEAM · EVERY ASSET edit, regenerate, approve OUTER LOOP · · BUILD, REVIEW, REFINE Ótimo sets up the design system, then reviews the output and updates the tokens and the Skill every month Brand system built in Figma Brief in plain words Generate Claude drafts it Review people decide Ship on-brand, signed off
The Brand OS loop Five steps in a column: brand system in Figma, brief, generate with Claude, review and ship. A dotted inner loop runs from review back to generate for your team on every asset. A gradient outer loop runs from review back to the brand system: Ótimo sets up the design system, then reviews the output and refines it every month. Brand system built in Figma Brief in plain words Generate Claude drafts it Review people decide Ship on-brand, signed off
  • Inner loopYour team, on every asset: edit, regenerate, approve.
  • Outer loopÓtimo sets up the . Every month: review the output, then update the tokens and the Skill.

Every asset, step by step.

Most of the effort sits with the system. The decision sits with a person.

01
Your team

Brief it

Ask for what you need in plain words, or start from one of your templates.

Claude
02
Claude

Generate

Claude drafts from your , and voice rules, so the first version already looks like you.

ClaudeFigma
03
A person decides

Review

Someone on your team approves it, edits it or sends it back. Anything making a regulated claim gets a named sign-off.

Figma
04
Your team

Ship

Export in the format the channel needs and publish.

Figma
05
Ótimo, monthly

Review and refine

Each month we review what the system made, turn your team's edits into rules, update the tokens and the , and test again.

ClaudeFigma

Claude takes the brief and drafts it, reading your brand straight from Figma, where it lives as tokens and components. Figma holds that brand from the draft through to the file you ship, and every month after.

What your brand can make.

Pitch decks, social sets, reports, proposals, emails and web pages, all generated from the same system and approved by your team before they go out.

The brief Your team

“Ten slides for the Q3 investor update. Lead with the new partnership and keep the numbers on one page.”

Written in plain words, or picked from a template.

First draft Claude

Drafted on system in minutes.

Shipped Approved

Reviewed, polished and signed off by a person.

The brief Your team

“A launch set for LinkedIn and Instagram: three posts with the same headline, sized for each channel.”

Written in plain words, or picked from a template.

First draft Claude

Drafted on system in minutes.

Shipped Approved

Reviewed, polished and signed off by a person.

The brief Your team

“A twenty-page annual impact report from last year's figures, in the house editorial style.”

Written in plain words, or picked from a template.

First draft Claude

Drafted on system in minutes.

Shipped Approved

Reviewed, polished and signed off by a person.

The brief Your team

“A proposal for a new financial services client, using our standard scope and pricing pages.”

Written in plain words, or picked from a template.

First draft Claude

Drafted on system in minutes.

Shipped Approved

Reviewed, polished and signed off by a person.

The brief Your team

“The monthly newsletter with four stories and one clear call to action.”

Written in plain words, or picked from a template.

First draft Claude

Drafted on system in minutes.

Shipped Approved

Reviewed, polished and signed off by a person.

The brief Your team

“A landing page for the autumn event, with speakers, the agenda and a sign-up form.”

Written in plain words, or picked from a template.

First draft Claude

Drafted on system in minutes.

Shipped Approved

Reviewed, polished and signed off by a person.

What stays human.

The system is only ever as good as the thinking behind it, and that part stays with people.

Kept with people

People decide

The system produces the work. It never decides what the brand is, or what is good enough to go out with your name on it.

  • Brand strategy and positioningWhat the brand is for, and who it is for.
  • Identity, naming and the big creative callsThe decisions the system is built on, not generated by it.
  • What's good enough to carry your nameEvery asset is approved by a person before it ships.
  • Anything that makes a regulated or high-stakes claimNamed sign-off, every time.

Gets sharper every month.

Every correction your team makes becomes a rule a designer has checked, so the next draft lands closer without picking up anyone's shortcuts.

CorrectionsWhat your team changed, collected each monthRulesA designer checks which edits should become a ruleTokensThe Figma source is updated, so the change is everywhereRepackaged, tested on real work, back in your hands

↺ Then it starts again next month.

This monthly loop is Brand OS Care. See what it costs

Questions about the loop.

Where does the human stay in the loop?
Throughout. A person owns the brand strategy and the system's setup, reviews and polishes the hero pieces, and signs off anything that makes a regulated or high-stakes claim before it ships. The AI handles the volume; a human keeps custody of the brand, deciding what's good enough to carry your name. You're never handing the brand to a machine, you're giving your team a faster way to make more of it, on-brand.
What don't you automate?
The thinking. Brand strategy and positioning, logo and identity creation, naming, and the big creative calls stay with people, because they need taste, judgement and a point of view a machine doesn't have. What we automate is the repetitive production downstream of those decisions: the decks, social, email and everyday assets that simply need to be on-brand and out the door. The system is only ever as good as the brand and the thinking behind it, and that part is human.
Does the AI actually stay on-brand?
It generates on-brand by construction. Claude reads your real and rules straight from the Figma source, so consistency is the default rather than something to police. Expect around 80% of everyday assets on-brand in minutes, with a person polishing the rest.

Plenty more on ownership, data and pricing.

See all questions