For brand teams
Fewer off-brand drafts to send back.
Agents read your rules first.
Brand compliance software built from your brand book: every finding cites its rule.
No exclamation marks: Lumo doesn’t shout.
“A wink, not a joke. No exclamation marks.” — Tone of voice, Lumo Drinks
- Headline ≤ 30 charactersMeasured
- Banned phrasesMeasured
- Claims and legalAI judged
No exclamation marks: Lumo doesn’t shout.
“A wink, not a joke. No exclamation marks.” — Tone of voice, Lumo Drinks
- Headline ≤ 30 charactersMeasured
- Banned phrasesMeasured
- Claims and legalAI judged
What changes for you
Brand compliance before review. Not after it.
-
One system, not ten PDFs
Positioning, tone of voice and guardrails become documents every sub-brand inherits, so nobody works from an old PDF.
-
Findings that cite the rule
Brand check reads the text of posts and visuals and quotes the passage behind each finding.
-
Changes wait for you
New files, competitor findings and comments arrive as tracked changes with a reason, and you keep or drop each one.
-
One brand in every AI tool
Claude, ChatGPT and other AI tools read your documents from Monobase, so nobody pastes the brand book again.
Where it happens
Four places you’ll use most. Each one opens its module.
- Lumo Drinks’ brand definition, ten of ten documents: Positioning and Tone of voice, with the tonality axes.
Your brand as documents
Ten documents agents read, from positioning to guardrails.
- Check rules drawn from Lumo’s Claims and legal document, each with its severity; one rule is still proposed.
Rules from your documents
The agent turns a document into rules Brand check applies.
- The Brand agent’s change proposal with its reason, the two sources it read and an assumption, with Keep and Discard.
Changes you approve
An agent’s edit arrives with its reason; you keep or discard it.
- Four of the read tools an AI client gets from Monobase, limited to one workspace.
The same brand in AI tools
Claude, ChatGPT and other AI tools read it, read-only by default.
What we don’t do
Brand check reads text first. Pictures come next.
Today it checks headlines, captions and copy against your documents. Checks of photos, logos, colours and layout are coming next; until then your templates and design system hold them. Documents have no named owner or review schedule yet.
See the rule behind every finding →Questions brand teams ask
Is Monobase brand compliance software?
Yes, starting with text. Brand check reads the copy of posts and visuals against rules drawn from your Brand System and cites the rule behind every finding. Checks of photos, logos, colours and layout, and of articles and newsletters, are coming next.
Does Brand check stop a post from going out?
Brand check warns; a person decides. Today a finding shows in the studio while you work. A second warning when someone approves or schedules a post with an open finding is coming next, and approving anyway is recorded.
Does Monobase learn from our feedback?
Monobase is starting to. Your team can already rate an agent’s answer with a reason. Next, themes that come up in several ratings or dismissed findings become a proposed change to your Tone of voice, which you approve before agents use it. No model is trained on your data.
Do we have to rewrite our brand book?
No. Drop it into the Brand agent’s intake as PDF, Markdown or plain text. The agent reads it, asks what is missing and drafts the documents; you review every change.
Can agencies and other teams change our rules?
Only through proposals. Owners and admins build the Brand System; creators work from it. Every agent change arrives as a tracked change a person keeps or drops.
Does Brand check give a score?
No. It lists the rules that were met and the findings to fix, each with the passage it came from and whether code measured it or AI judged it.
What about sub-brands and products?
Each level of your structure holds its own documents. Binding documents apply to every level below, so a product inherits its brand’s tone of voice.
Where is our data?
Database, app and agents run in Frankfurt, and owners can switch on EU-only mode for AI models. The security page has the details.