Forbidden phrases caught at every AI output — competitors, off-brand words, regulated claims
Catch competitor names, off-brand language, regulated claims, and deprecated terms at every AI output — pattern matching plus AI semantic detection across your whole content stack.
The problem
You have a banned-words list in a Google Doc — competitors you do not mention by name, deprecated product terms, off-brand slang, claims that legal does not allow. Your AI ignores the doc. Generic content moderation tools handle user-generated content. They do not know your brand's specific forbidden phrases.
You tried OpenAI's Moderation API, but it catches hate speech and harassment — not "do not use the word revolutionary in our copy." Google Perspective and Hive cover similar ground at the platform level. Built for forum, gaming, and social user-generated content at scale, not for brand-produced AI content with your specific prohibitions.
Profanity filter APIs (Bad Words API, ProfanityCensor, WebPurify at $10 to $1,000 per month) handle generic profanity but require custom integration plus per-brand list maintenance. Content moderation services (Telus International, TCS, Genpact at $0.10 to $2 per item) put humans in the loop for user-generated content — wrong layer for brand-produced AI content.
Your Google Doc goes stale within a quarter. Competitor names slip into comparison content. Deprecated product names appear in new location pages. Regulated claims appear without substantiation. Trademark-protected competitor language shows up in paid creative.
What success looks like
Every AI-produced content output gets checked against a brand-specific forbidden phrase library before it publishes. Competitor names get caught. Off-brand language ("guys" instead of "team," "revolutionary" when the brand avoids hyperbole) gets caught. Regulated claims requiring substantiation get blocked unless the substantiation is on file. Deprecated product terms get blocked. Trademark or IP-protected competitor language gets blocked.
Pattern matching catches exact-string prohibitions cheaply and fast. AI semantic detection catches paraphrased violations pattern matching cannot find — synonyms of off-brand language, implied prohibited claims.
Multi-brand portfolios and multi-industry operators get per-brand and per-industry library extensions with an inheritance hierarchy. Adds and removes flow through review across corporate, franchisee council, legal, and compliance.
The library works alongside the broader brand voice check and the marketing compliance check. Together they enforce every prohibition you have defined at every output — location pages, Google Business Profile, review responses, social, email, paid creative, product descriptions, support replies.
How most operators solve this today
A few categories of moderation tools exist. None of them enforce your specific brand prohibitions on AI-produced content:
AI content moderation (Hive, Spectrum Labs, Two Hat/Microsoft, OpenAI Moderation, Google Perspective, Modulate)
Free to $50,000+/month
Built for user-generated content moderation (forums, gaming, social) at scale. Catches hate speech, harassment, sexual content, violence. Does not enforce your specific brand prohibitions.
Profanity filter APIs (Bad Words API, ProfanityCensor, WebPurify)
$10 to $1,000/month
Generic profanity detection. Requires custom integration plus per-brand list maintenance. No semantic detection. No review workflow.
Content moderation services (Telus International, TCS, Genpact)
$0.10 to $2 per item
Humans in the loop for user-generated content. Wrong layer for brand-produced AI content.
CMS profanity plugins (WordPress, Drupal modules)
$10 to $100/year
Bolted into your CMS. Generic profanity filtering. Does not extend to AI producing content outside the CMS.
In-house (regex blocklist in code + Google Doc + brand manager review)
Internal time
The Google Doc goes stale within a quarter. No semantic detection. No integration with your AI. No audit trail.
Build it in-house
Senior engineer ($130-220k) + brand manager time
Regex matching is easy. The semantic detection that catches paraphrased violations, the multi-brand inheritance, and the review workflow are the parts that take quarters to build.
What changes when this is an agent skill
Combines pattern matching and AI semantic detection to check every AI-produced output against your forbidden phrase library before it publishes. Pattern matching catches exact-string prohibitions cheaply and fast. The AI pass catches paraphrased violations — synonyms of off-brand language, implied prohibited claims, near-misses regex cannot find.
The library is yours, brand-specific. Competitor names you do not mention. Deprecated product terms. Off-brand slang. Regulated claims requiring substantiation. Hyperbole or superlatives the brand avoids. Trademark or IP-protected language. Plus the industry-specific prohibitions that regulators care about.
Multi-brand portfolios and multi-industry operators get per-brand and per-industry library extensions. Corporate sets the base prohibitions. Brand and industry sub-libraries extend or override within bounds. Every add and remove flows through review (corporate, franchisee council, legal, compliance) with audit trail.
The library works alongside the broader brand voice check and the marketing compliance check. Together they catch every prohibition you have defined at every AI output — location pages, Google Business Profile, review responses, social, email, paid creative, product descriptions, support replies.
The total cost replaces the Google Doc plus the brand manager review time plus the per-platform profanity filter subscriptions.
Agents that include this skill
Skills live inside agent rentals. To get this skill in production, hire any of the agents below — context-tuning at onboarding is included in the first month.
Brand-Spec Authoring + Maintenance Agent
Produces and maintains the canonical brand spec every content-producing agent's brand-voice gate enforces.
FAQ
- What does the forbidden phrase library actually do?
- Holds a list of phrases you do not want appearing in AI-produced content — profanity, competitor names, off-brand slang, regulated claims without substantiation, deprecated product terms, trademark-protected competitor language — and checks every output against the list before it publishes.
- How is this different from OpenAI Moderation or Google Perspective?
- Those detect hate speech, harassment, sexual content, and violence in user-generated content. This catches your specific brand prohibitions (competitor names, off-brand language, regulated claims, deprecated product terms) in brand-produced AI content.
- How is this different from a content moderation service like Telus or TCS?
- Those use humans to moderate user-generated content at $0.10 to $2.00 per item. This catches violations in AI-produced content automatically at the moment of publishing.
- What kinds of phrases does the library include?
- Competitor names. Deprecated product terms. Off-brand slang. Regulated claims requiring substantiation. Hyperbole or superlatives the brand avoids. Trademark or IP-protected language. Profanity. Industry-specific prohibitions from your regulators.
- How does the AI semantic detection work?
- A classifier catches paraphrased violations the pattern match cannot — synonyms of forbidden terms, implied prohibited claims, near-misses. Runs in combination with the pattern match for efficiency.
- Can different brands or industries have different libraries?
- Yes. Multi-brand portfolios get per-brand libraries. Multi-industry operators get per-industry extensions. Corporate sets the base prohibitions. Brand and industry sub-libraries extend or override within bounds.
- How does this work with the brand voice check and the marketing compliance check?
- The brand voice check is the broader check that scores tone, formality, and structure. The marketing compliance check handles regulator-required prohibitions. This holds the brand-specific forbidden phrase list. All three run at the moment of publishing.