Brand voice review

This playbook defines an automated, human-supervised brand voice review for one submitted content draft, such as a blog post, newsletter issue, announcement, or landing page draft. The run begins when a writer or editor submits a candidate piece along with its target channel and audience. It evaluates the text against company-supplied voice pillars, tone spectrums, approved copy examples, and banned vocabulary lists stored in connected workspaces such as Notion, Google Docs, or CMS platforms. The agent operates strictly within an editorial review and feedback boundary: it flags off-brand language, prepares grounded rewrite suggestions, and coordinates revisions with human editors. It does not publish live articles, send customer emails, modify production CMS statuses, or execute marketing campaigns.

• Retrieve the candidate draft, metadata, target channel, and designated author from the submitted case payload or source document in Notion, Google Docs, or CMS repository. If the source text is empty, incomplete, or behind inaccessible permissions, pause the case and request the accessible document link or raw text from the submitter before proceeding.
• Retrieve the company's approved brand voice documentation, including core voice attributes, tone guidelines by channel, positive and negative copy examples, and lists of prohibited phrases or competitor terminology. If voice reference documentation is missing or ambiguous regarding the target channel, pause the case and prompt the brand lead to provide the applicable rules.
• Segment the draft into logical sections and compare each paragraph against the established voice attributes and tone spectrums. Identify specific instances of passive voice, corporate jargon, ungrounded superlatives, overly casual colloquialisms, or misaligned emotional tenor that deviate from documented brand standards.
• Cross-reference all proper nouns, product feature descriptions, competitive claims, and compliance-sensitive terms against the company's approved vocabulary and claims rules. Flag any prohibited terminology, unapproved feature names, or unverified performance guarantees that lack supporting substantiation.
• For each identified discrepancy, construct an evidence-backed critique citing the specific governing brand rule, contrasting positive and negative examples from the brand library, and generating a proposed alternative rewrite that preserves the author's original factual intent while aligning with brand tone.
• Information request. When a draft introduces novel terminology, product metaphors, or positioning angles not addressed in existing brand guidelines, ask the brand lead whether the new framing is acceptable brand evolution or must be revised back to documented standards. Present the extracted terminology, contextual sentences, and proposed alternative phrasings.
• Compile the complete voice evaluation into a structured editorial packet containing overall voice alignment scores, itemized line critiques, side-by-side alternative rewrites, and any outstanding terminology questions.
• Approval. Present the compiled voice review packet to the designated brand editor or content lead for review. Request the editor to approve the suggested revisions, provide custom alternative phrasing, or return the draft to the author with itemized feedback. Include the full diff, rationale, and governing brand guidelines in the review context.
• If the brand editor requests revisions or provides custom phrasing, apply the approved modifications to create an updated candidate draft. Share the revised text with the original author for confirmation, verifying that factual meaning and core technical points remain intact.
• Exception. If the author or brand editor rejects the voice critique and cannot agree on acceptable phrasing, escalate the impasse to the marketing director with the competing perspectives and brand guideline citations to render an authoritative editorial decision.
• Once all approved voice edits are accepted by the author and brand editor, record the finalized text, sign-off timestamps, and review audit trail in the tracking system. Notify the content team that the draft has achieved editorial voice acceptance and is ready for separate downstream formatting and staging.
• If the draft cannot achieve voice alignment due to fundamental scope mismatch, unresolvable claims violations, or author withdrawal, mark the case declined or unresolved, document the specific blocking criteria and review history, and close the case without sign-off.
• A freelance writer submits a product launch article for a B2B SaaS platform describing the software as an "unbeatable, revolutionary AI disruptor that completely eliminates customer churn." The agent evaluates the draft against the company's tone pillars in Notion, which mandate pragmatic humility, customer-centered problem solving, and zero hyperbolic superlatives. The agent notes that "unbeatable" and "revolutionary" violate explicit voice rule #3 (no ungrounded superlatives), and "completely eliminates customer churn" violates product marketing compliance rule #8 (no absolute outcome guarantees). The agent drafts a replacement sentence: "Our automated workflow helps customer success teams identify at-risk accounts early and systematically address churn drivers." The brand editor reviews the comparison packet, approves the suggested rewrite, and prompts the writer. The writer accepts the change, and the piece is stamped voice-accepted.
• The case is complete when either the brand editor and author accept the finalized draft and its audit trail is recorded, or the case is documented as declined or unresolved due to insurmountable voice violations or withdrawal. An accepted review confirms editorial voice clearance only; downstream publishing, scheduling, and distribution remain separate team actions.