Examples
Example 1: Blog Post Blocked for Voice Drift and a Prohibited Phrase
A SaaS client's account manager has asked for a thought-leadership post on their product's approach to data security. Marketing Content finishes the draft and the account manager triggers publishing.
Step 1: Automatic audit. The publish flow sends the draft to Brand Guard. Brand Guard loads the client's brand context: a measured, evidence-first voice register and a rules set that prohibits "guaranteed ROI" and "cutting-edge."
Step 2: Scoring. Brand Guard runs its checks. The opening paragraph uses "In a world where cutting-edge security solutions are table stakes..." which triggers a prohibited-phrase flag and an AI-signature flag. The body scores well on Messaging and Governance but Voice lands at 19 out of 30 because the register is more generic than the client's usual style.
Step 3: Block notice. A one-line notice appears in the Marketing Content channel: "Blocked draft: score 77 / Voice 19. Prohibited phrase (cutting-edge), AI-signature opener. See the card for redlines."
Step 4: Redlines. The Marketing Content card shows two redlines. First: the opening line, flagged as a prohibited phrase and AI-signature, with a proposed replacement leading directly with the client's specific security claim. Second: a mid-body paragraph that uses "leverage" twice, with a proposed rewrite using plain language.
Step 5: Revise and resubmit. The account manager applies both edits and resubmits through Marketing Content. Brand Guard scores the revised draft at 91: auto-approved. The post moves to the client review queue.
Example 2: LinkedIn Post Auto-Approved, One in a Batch Flagged
Marketing Social generates five LinkedIn posts for a B2B professional-services client, covering three of their content pillars. The account manager triggers publishing for all five.
Step 1: Batch audit. Brand Guard scores each post against the client's brand rules. Four score between 87 and 93 and clear automatically. The fifth scores 74 because its closing CTA ("Reach out to learn more") is generic and the client's brand rules require a specific offer in every CTA.
Step 2: Partial auto-approval. The four passing posts move forward without any action needed. A one-line notice appears in the Marketing Social channel for the fifth: "Blocked draft: score 74 / Messaging 17. CTA does not meet per-client specificity rule. See card for redlines."
Step 3: Single redline. The card shows one redline: the CTA paragraph, with a proposed replacement that names the client's specific offer ("Book a 30-minute attribution walkthrough") in place of the generic "Reach out."
Step 4: Quick fix. The account manager accepts the proposed CTA, resubmits, and Brand Guard scores the revised post at 89. All five posts are now approved.
Example 3: Conversational Check Before a High-Stakes Draft
Before starting a formal draft, an account manager wants to know whether a particular angle will clear brand review for a financial-services client. They paste a rough paragraph into the brand-guard channel.
Step 1: Direct question. The account manager writes in the brand-guard channel: "@Agenteous look at this paragraph for [client]: the angle is that their approach to risk management is more conservative than their competitors. Would this pass?"
Step 2: Brand Guard's response. Brand Guard loads the client's brand context. The client's governance rules flag any comparative claim about competitors unless the claim is documented and sourced. Brand Guard replies: "The angle is on-pillar for this client, but the phrase 'more conservative than their competitors' triggers the comparative-claims rule. Either name the specific comparison with a source, or reframe as 'the firm's documented approach to X.' Voice and Messaging look strong otherwise."
Step 3: Pre-flight adjustment. The account manager adjusts the brief before sending it to Marketing Content, so the draft is constructed correctly from the start rather than blocked after the fact.
Example 4: Operator Overrides a Block
Marketing Email sends a re-engagement campaign draft for a client. Brand Guard scores it at 68 and issues a revise verdict, flagging that the subject line uses "Re:" as a prefix, which the client's governance rules prohibit.
Step 1: Block and redline. A one-line notice appears in the Marketing Email channel. The card shows one redline: remove the "Re:" prefix from the subject line.
Step 2: Operator decision. The account manager reviews the redline and realises the "Re:" was intentional: the client specifically requested it for this campaign to match a test they are running. The account manager clicks "Publish now" in the Marketing Email card.
Step 3: Override recorded. Brand Guard logs the override, the operator who approved it, and the campaign reference. The email is queued for sending. The override appears in the audit history so it can be reviewed in the next retrospective and, if appropriate, a per-brand rule exception can be configured for future sends.