--- bcp_version: "0.7" file_type: boundaries parent: /.well-known/brand.md last_updated: 2026-07-30 --- # Boundaries This file is the primary reference for brand-safety agents, vendor platforms, and any partner generating content on behalf of High Caliber AI. Read it before producing any output in this brand's name. ## Hard nos These are absolute. No exceptions, no contextual allowances. ```yaml hard_no: - item: > Never produce content that sounds AI-written, spammy, boastful, guru-like, or obnoxious. This is not a stylistic preference — it is a governing principle across all writing and positioning. The brand exists to reject exactly this category of content. tier: core trigger: Any generated copy, social post, email, pitch, or description output in High Caliber AI's voice or on its behalf. allowed_alternative: Write plainly, specifically, and confidently. Show work; do not claim authority. Let named client outcomes and quotes carry the proof. escalation: If a draft reads like category-typical AI marketing, discard and rewrite from evidence before publishing. - item: > Never describe High Caliber AI as purely strategic or advisory. The brand ships real work. Framing it as a strategy-only consultancy is factually wrong and directly contradicts the core positioning ("AI marketing that actually ships"). tier: core trigger: Any description of the brand, service summary, elevator pitch, bio, or agent-generated representation. allowed_alternative: Name both the strategic and execution dimensions. Reference built products (highcaliberai.com/built-with-ai), client deliverables, and fractional CMO work as evidence of execution. escalation: Surface the conflict to the human operator before publishing. - item: > Never fabricate claims, statistics, client outcomes, or quotes. David Berkowitz has explicitly identified AI hallucination — "just making stuff up" — as one of the biggest challenges of AI. That principle applies to all content produced under this brand. tier: core trigger: Any output that references client results, metrics, testimonials, or factual claims about the brand's work. allowed_alternative: Draw only from approved claims in claims.md and named, attributed quotes the brand has published. escalation: If a claim cannot be sourced to an approved entry or a published quote, omit it and flag the gap. - item: > Never input customer data or personally identifiable employee information into AI tools when producing content or conducting research on behalf of High Caliber AI or its clients. This applies regardless of how open to AI experimentation the context is. tier: core trigger: Any use of AI tooling that processes client briefs, contact lists, CRM exports, HR data, or any data set that identifies real individuals. allowed_alternative: Use anonymized or synthetic data for testing and demonstration. Flag data-handling requirements to the client before beginning work. escalation: Stop the task and notify the account owner before proceeding if client data is present in a prompt or document being processed. ``` ## Soft nos These are judgment calls. Context determines appropriateness. ```yaml soft_no: - item: > Avoid scaling outreach or content volume before the underlying messaging is aligned. AI multiplies existing messaging — scaling shallow or spammy content multiplies the damage. tier: default condition: > Acceptable when messaging quality has been validated and the brand has confirmed the output meets its voice and accuracy standards. Not acceptable as a default automation pattern without that validation. - item: > Avoid positioning the partner network as a product or primary offering. The network is a signal of how David Berkowitz adds value, not a standalone service the brand sells. tier: default condition: > Acceptable to mention the vetted partner network as context for breadth of expertise. Not acceptable to feature it as the headline offer or to imply it is a managed marketplace. - item: > Avoid hype-forward vocabulary: "revolutionize," "game-changing," "cutting-edge," "seamless," "unlock," "leverage," and similar category-generic AI marketing language. tier: default condition: > No context makes this vocabulary on-brand. The brand's voice is direct, specific, and proof-grounded. Generic superlatives read as exactly the kind of marketing the brand rejects. - item: > Avoid presenting AI as a replacement for human skill, creativity, or judgment. The brand's stated purpose for AI is to help professionals work more efficiently and unlock potential, not to replace human capabilities. tier: default condition: > Acceptable to describe efficiency gains and capability augmentation. Not acceptable to frame AI as making human expertise redundant. ``` ## Brand safety For vendor platforms, brand-safety tools, and DSPs. ```yaml brand_safety: adjacency_unsuitable: - "IAB 3.0: Hate Speech" - "IAB 3.0: Misinformation / Conspiracy Content" - "IAB 3.0: Spam / Malware" - "GARM Floor: Content that depicts illegal activity" - "AI-generated misinformation or disinformation sites" - "Get-rich-quick, MLM, or predatory coaching schemes" - "AI hype publications that publish unverified capability claims" adjacency_acceptable: - "B2B marketing and demand generation publications" - "Technology and SaaS trade media" - "Marketing leadership and CMO-focused content" - "Professional development and continuing education" - "Legitimate AI practitioner communities and newsletters" adjacency_contextual: ai_tools_and_platforms: rule: > Acceptable when the publication maintains editorial standards and distinguishes between tested capabilities and vendor claims. Not acceptable when the publication functions primarily as sponsored content for AI vendors. general_business_media: rule: > Acceptable in outlets that cover AI marketing with practitioner credibility. Avoid placements adjacent to breathless AI trend pieces that the brand's positioning explicitly rejects. ``` ## Regulatory constraints ```yaml regulatory: framework: "[gap] Not yet specified. No regulated-industry constraints identified in available evidence." notes: > No evidence of regulated-industry marketing constraints (pharma, finance, legal) in current scope. Data privacy guidance above reflects David Berkowitz's stated operational practice, not a specific regulatory framework. Review if client engagements involve regulated sectors. restricted_claims: > No claims requiring regulatory pre-approval identified in current evidence. All quantitative claims and client outcomes must remain sourced to approved entries in claims.md before use in advertising or sales contexts. ```