What is the Brand Brain?
The Brand Brain is the memory a Digital FTE holds about the business it works for: what the company does and for whom, its products and their facts, how it speaks and which words it avoids, what it claims and what it never claims, and the rules its manager has set. It lives in the Memory tab of the workspace, and every item in it shows its provenance, meaning where the FTE learned it.
The name reflects the role it plays: the FTE's working knowledge of the brand, held where a person can read it, rather than a hidden store that shapes drafts invisibly.
How is the Brand Brain built and kept?
It is built in four ways, each labelled. From onboarding, when the organization told the platform what it is. From documents the manager provides, such as brand guidelines and product documentation. From connected systems, such as the product catalogue in the CMS. And from inference, when the FTE draws a rule from a pattern in the manager's approvals and send-backs, such as never mentioning a competitor by name after several send-backs for doing so.
The manager reviews it fully at hiring in the first-run wizard, corrects items as proposals reveal them wrong, and confirms the whole memory quarterly when the workspace prompts. Inferred items are shown as inferred so they can be confirmed or replaced with explicit rules.
Why does provenance matter?
Provenance matters because it tells the manager how much to trust an item and where to fix it if it is wrong. An item from onboarding is something a person typed; one from a document can be traced to the document; one inferred from approvals is a generalisation that may be right or may be too broad. Without provenance a memory is a black box; with it, every belief the FTE holds can be checked.
It also makes corrections precise. A wrong product fact traced to an outdated document is fixed by replacing the document; a wrong rule traced to inference is fixed by writing the explicit rule; and the fix propagates to every later draft.
How is it shared, and how is it isolated?
The Brand Brain is per property. A Search Visibility FTE managing two websites holds a separate memory for each, and a correction made on one never touches the other. Within a property it is shared across roles: a fact corrected once is corrected for the Content FTE, the Social Growth FTE, and any other role working on that property.
That combination is what lets an agency run several clients from one organization without cross-contamination, and what lets a multi-role team make each correction once.
- Per property: two websites, two memories, no leakage
- Shared across roles on a property: one correction, every FTE
- Confirmed quarterly: businesses change quietly
- Never holds personal data about individuals
What is the Brand Brain often confused with?
It is confused with a model's training, which it is not: the platform does not train models on customer data, and the Brand Brain is a store of confirmed items the FTE reads, not weights it has absorbed. It is confused with a knowledge base for customers, which is a source the Support FTE cites; the Brand Brain is the FTE's own understanding of the business, and the knowledge base is one of the documents it may draw on.
It is sometimes assumed to fill itself in. It is populated from the four provenances, and anything the FTE cannot source stays out of it; the manager's review is what makes it reliable, not the FTE's confidence.
Where do you see it in the workspace?
The Memory tab shows every item grouped by kind, each with a provenance chip and a confirmed or unconfirmed state, with a quarterly confirmation prompt. Changes are logged. Because the Brand Brain is a source for the fact gate, a wrong item there produces sourced-but-wrong claims, which is the case the gate cannot catch and the review at hiring exists to prevent.
Related terms: fact gate, property, remit, approved sources, quarterly confirmation.