Brand Consistency Checker

Brand Consistency Checker

Purpose

Acts as the brand quality gate for all Synthex content. Scores submitted content against the Business DNA profile across five brand dimensions, flags specific issues, and returns actionable rewrites for any off-brand elements.

This turns Synthex into a brand compliance tool — ensuring every post published through the platform sounds like the brand, not like a generic AI output.

Workflow

1. Receive content to audit (paste, post ID, or scheduled post reference)
2. Load Business DNA for this brand/business
3. Score across 5 brand dimensions
4. Flag specific violations with line-level detail
5. Suggest targeted rewrites for flagged sections
6. Return PASS / CONDITIONAL / FAIL verdict

Brand Scoring Dimensions

Score each dimension 0–100. Overall = weighted average.

1. Tone of Voice (30%)

Compare the content's tone against dna.toneOfVoice and dna.languageStyle.

Check for:

• Formality level match (formal vs casual)
• Energy level (enthusiastic vs measured)
• Perspective (first-person brand voice vs third-person)
• Authenticity (genuine vs generic "AI-sounding" phrases)

Automatic fails (score 0):

• Generic opener: "In today's fast-paced world..."
• Hollow superlatives: "world-class", "cutting-edge", "game-changing" (unless in brand vocabulary)
• Passive voice overuse (>20% of sentences)

2. Vocabulary Compliance (25%)

Check dna.vocabulary (required terms) and dna.avoidWords (forbidden terms).

Required terms present: [count]/[total] ✓/✗
Forbidden terms found: [list any violations]
Industry jargon: appropriate/excessive
Competitor references: none/flagged

3. Message Alignment (20%)

Compare key messages against dna.usp, dna.keyBenefits, and dna.differentiators.

• Does the content reinforce the USP?
• Are key benefits represented?
• Is the target audience (dna.targetAudience) addressed?
• Is the BYOK/value proposition present (if relevant)?

4. Platform Appropriateness (15%)

If platform is known, check against platform rules in platform-content-adaptor skill:

• Length within platform limits
• Hashtag count appropriate
• Format correct (thread vs single, portrait vs landscape reference)
• CTA present and appropriate to platform

5. CTA and Conversion (10%)

• CTA present: yes/no
• CTA matches campaign goal: yes/no/partial
• CTA is specific and actionable (not "learn more" alone)
• Urgency present (if promotional content)

Scoring Thresholds

Output Format

──────────────────────────────────────────
 BRAND CONSISTENCY AUDIT
 Brand: [name]  |  Platform: [platform]  |  Date: [DD/MM/YYYY]
──────────────────────────────────────────

OVERALL SCORE: 76/100 ⚠️ CONDITIONAL

DIMENSION SCORES
  Tone of Voice:         82/100 ✓
  Vocabulary:            90/100 ✓
  Message Alignment:     65/100 ⚠
  Platform Fit:          80/100 ✓
  CTA & Conversion:      60/100 ⚠

ISSUES FOUND

  [Line 3] "In today's competitive landscape..."
  → OFF-BRAND: Generic opener. Your brand voice is direct and specific.
  → SUGGESTED: "[Brand name] solves [specific problem]..."

  [Line 8] Missing: USP reference. Your USP is "[usp]"
  → Add a line that connects this post back to your core offer.

  No CTA detected.
  → Add: "[Your CTA from Business DNA]"

VERDICT: ⚠️ CONDITIONAL — fix 3 issues then publish.

Integration with Synthex Workflow

• Pre-publish gate: Run this check before approving any AI-generated post
• Batch audit: Check all scheduled posts in a campaign before launch
• Persona training: Patterns from failing posts improve persona prompts over time

Use lib/ai/content-scorer.ts for the Platform Fit and Engagement dimensions. Apply brand dimension scoring via LLM call using the Business DNA as context.

Reference

• Business DNA: .claude/skills/business-dna/
• Content scorer: lib/ai/content-scorer.ts
• Persona system: lib/ai/content-generator.ts (personaId + system prompt)
• Brand profile: app/dashboard/settings/brand-profile

Capability Uplift — Override Defaults

NEVER output vague brand feedback: "this sounds on-brand", "good feel", "consistent with your voice", "aligns with your brand values". This feedback is useless — it cannot be acted on. Never approve content that contains anti-pattern phrases from content-standards.md without flagging them.

INSTEAD every consistency check produces a structured score:

BRAND CONSISTENCY REPORT ───────────────────────── Vocabulary match: [X]% — [N] DNA words used, [N] missing from expected set Anti-patterns found: [N] — [list each phrase explicitly] CTA quality: [Pass/Fail] — [exact CTA text] → names action: [Y/N] + benefit: [Y/N] Tone match: [X]/10 — [specific description of what's aligned and what isn't] Overall: [Pass ≥7/10 / Revise <7/10]

Specific fixes:

• Line [N]: Replace "[exact phrase]" → "[suggested replacement]"

REFERENCE .claude/skills/synthex-standards/references/content-standards.md

Foundation & Gate Wiring (SYN-1049)

Adopted from the senior-skill standard so every artefact this connector produces is checked against the locked foundation before it lands.

Reads at every invocation (never cached — re-read each run):

• .claude/memory/ceo-foundation.md — voice tag (Q2.5.5), Phase 4 voice amendments, universal + brand-specific taboos (Phase 1 / Phase 3.X).
• .claude/memory/verification-gates.md — gate state for any claim referenced.

Output gate: every client-facing artefact this connector produces routes through brand-voice-enforce before the CEO batched-review queue. A REJECT blocks the artefact until the quoted offending string is fixed.

Evidence standard: every quantitative or factual claim carries exactly one tag — [VERIFIED] / [INFERENCE] / [UNCONFIRMED]. Untagged = defect (.claude/rules/fabel-evidence-standard.md). Never state a projected result as fact.

Spec: see spec.md in this skill directory.