What a brand hub should hold
A PDF brand book describes the brand; a brand hub has to hold it in a form tools can use. That means logo files in vector form, with primary, secondary, mono and reversed versions, minimum sizes and clear space; colours with values for each medium, HEX and RGB for screens, CMYK and spot references for print; typefaces with weights, fallbacks and licence terms; image style, including lighting, composition, casting and what never to show; tone-of-voice rules; approved claims with the markets where each may run; banned words; legal lines per market; and the rules for sub-brands and co-branding with partners.
Enforce at generation, not in review
Checking afterwards means reviewers catch deviations one asset at a time, inconsistently, and at AI volumes they cannot keep up. Enforcement at generation constrains the output from the start: the palette is fixed, type is set in the brand fonts, the logo is placed from the real file within its rules, tone rules and banned words shape the copy, and claims come only from the approved library. Automated checks still run on the result to catch what slips through, but review can then concentrate on judgement: whether the idea is right, whether the image feels like the brand, whether the market will read it as intended.
Guidelines as data
A rule a machine can enforce has to be precise. A logo at least 24 pixels wide on screen is checkable; use the logo sensibly is not. Text contrast of at least 4.5:1 for body text and 3:1 for large text, the WCAG 2.2 AA thresholds, can be tested on every asset. Headlines in sentence case, a minimum gap between logo and headline, never a logo over a face: each becomes a rule with a scope, such as a market, channel or sub-brand, and documented exceptions. Converting a brand book this way also exposes its contradictions, which is worth doing before any tool starts applying it.
Audit and change control
Guidelines change: a refresh adjusts the palette, legal withdraws a claim, a market adds a mandatory line. Each change should be versioned, with who made it, when and why, and every asset should record which version of the rules it was generated under. That record turns a withdrawn claim from a folder-by-folder trawl into a search: find every asset that used it, in every market, and replace it. Synthetic White's brand hub holds logo, colours, typography, image style, tone of voice, approved claims and banned words and enforces them at generation, with approvals, sign-off and an audit trail on every asset.
Updated 25 September 2026