Tester Quality

• Reviewing AI-generated copy for brand voice alignment
• Running pre-send checks on SMS, push, or in-app messages
• Diagnosing why AI Content QA flagged or passed a piece of content
• Configuring or updating Brand Guidelines to enforce tone and personality rules
• Advising on which channels support AI-assisted quality checks

Topics Synthesized

Brand Guidelines

Brand Guidelines in Braze's AI Copywriting Assistant allow teams to encode brand voice, tone, and personality constraints so that AI-generated copy stays consistent with organizational standards. Guidelines can be created from the message composer or via Settings > Brand Guidelines.

Key validation considerations:

• Are brand guidelines currently active and applied to the relevant campaign or canvas?
• Does the generated copy reflect the configured tone (e.g., formal, playful, authoritative)?
• Are prohibited terms, required phrases, or stylistic constraints being respected?
• When copy diverges from brand voice, check whether guidelines need updating or whether the AI prompt needs refinement

Compliance before launch: Brand Guidelines are not a post-publish audit tool — they are a pre-send configuration that shapes generation. If content feels off-brand at review time, the root cause is typically missing or under-specified guidelines, not a generation failure.

AI Content QA (BrazeAI)

AI Content QA is a pre-send content quality check powered by BrazeAI. It is available for SMS, Android push, iOS push, and traditional in-app messages only — it does not apply to email, Content Cards, or webhook steps.

What AI Content QA checks:

Validation lens — before launch:

• Run AI Content QA as a final gate before campaign activation
• A passing QA check indicates structural and tonal correctness; it does not guarantee conversion or engagement
• A flagged check should be treated as a blocker unless the team explicitly accepts the deviation with documented rationale
• QA results are advisory on unsupported channels — do not assume silent pass means implicit approval

Perspective: Content Correctness and Brand Compliance Before Launch

This skill approaches all content questions from the perspective of a pre-launch gatekeeper, not a creative collaborator. The goal is to catch problems before they reach end users.

When evaluating content:

1. Correctness first — Is the message factually and grammatically sound? Does it say what was intended?
2. Brand compliance second — Does it match the voice, tone, and constraints defined in Brand Guidelines?
3. Channel fit third — Is it appropriate for the delivery surface (character limits, push norms, SMS regulations)?
4. AI QA result last — Use the BrazeAI check as a final confirmation, not the first filter

If any layer fails, the content should not launch until the issue is resolved or formally accepted.

Common Failure Patterns

• Brand Guidelines not applied: AI-generated copy passes QA structurally but sounds off-brand — check that guidelines are linked to the workspace or campaign
• QA run on unsupported channel: AI Content QA is unavailable for email and Content Cards; teams sometimes assume a silent result means passing
• Flagged QA overridden without review: QA flags dismissed as noise rather than investigated — each flag should have a documented decision
• Guidelines too vague to enforce: If brand guidelines only say "be friendly," the AI has no actionable constraint; effective guidelines include concrete examples of in-voice and out-of-voice copy