beamix-voice-canon

Beamix Model B voice canon: when to name agents vs when to say 'Beamix', the no-AI-disclosure rule, banned buzzwords, and Hebrew+English dual-language parity expectations. Use for any copy, email, UI

SKILL.md

GitHub

Beamix Voice Canon — Model B

Quick reference

Agent names in product UI; Beamix on external docs. No AI disclosure. Three-word personality: Authoritative, Direct, Warm. HE + EN parity always.

When to use

  • Writing any user-facing copy: UI labels, email templates, onboarding text, PDF reports
  • Reviewing content for brand compliance before shipping
  • Configuring agent names shown in the product dashboard
  • Writing HE (Hebrew) or EN (English) copy for any surface

When NOT to use

  • For internal agent-to-agent communication (no voice canon required there)
  • For code comments (technical comments are direct and unbranded)

The three-word personality

Authoritative. Direct. Warm.

Every sentence should be checkable against these three: Does it sound like a confident expert? Does it get to the point? Does it feel human, not corporate?

Model B: named agents in product, "Beamix" on documents

Surface Use Example
Product UI — agent hub, job cards, action buttons Agent names "Shai is scanning your competitors"
In-product notifications, tooltips, status messages Agent names "Ran finished your FAQ update"
Emails, transactional and marketing "Beamix" "Beamix found 3 new ranking opportunities"
PDFs, reports, exports "Beamix" "Prepared by Beamix — {date}"
Onboarding completion seal "Beamix" "— Beamix" (not "— your crew")
Framer marketing site "Beamix" All marketing copy uses "Beamix"
Legal, terms, billing "Beamix" Full company name where required

The distinction: product = agents have names and personality. External documents = Beamix as the company.

No AI disclosure on content

Agent-generated content must read as human-written. Do NOT add:

  • "Generated by AI"
  • "Written by our AI agent"
  • "Powered by Claude"
  • Any disclosure label on content items

Adam handles disclosure decisions. Agents produce clean content with no labels.

Banned buzzwords and phrases

Never use these — not in copy, emails, PR descriptions, or doc titles:

leverage          enable            synergy           robust
seamless          cutting-edge      game-changing      next-level
innovative        best-in-class     world-class        ecosystem
holistic          empower           unlock             transform
journey           solution          space               utilize

If you catch yourself writing any of these, replace with a specific claim.

Banned Replace with
"leverage your AI visibility" "improve how AI search engines rank you"
"robust scan engine" "scans ChatGPT, Gemini, and Perplexity in under 60 seconds"
"seamless onboarding" "you're scanning in 3 steps"

Tone examples

Correct

Your GEO score dropped from 72 to 58 this week.
ChatGPT stopped mentioning you when people search for accountants in Tel Aviv.
Ran wrote 4 FAQ pages that address the exact queries where you're invisible.

Incorrect

Your AI search visibility has experienced a decline in performance metrics.
Leverage our robust agent system to seamlessly enhance your digital presence.
Our cutting-edge AI enables you to unlock your potential in the AI search ecosystem.

Hebrew + English parity

Beamix serves Israeli businesses as primary audience, with global secondary. Every user-facing surface must support both HE and EN.

Parity rules:

  • EN and HE copy must carry the same meaning — no shortcutting the HE version
  • HE is RTL — check that UI layout handles dir="rtl" correctly for HE locale
  • Idioms: translate the meaning, not the words. "Invisible to AI search" → HE equivalent carries the same market-positioning meaning
  • The no-buzzword rule applies to HE as well — find the direct Hebrew phrase

Taglines (canonical):

EN: "Stop Being Invisible to AI Search"
HE: "השותף החכם שעושה את השלבים בשבילך"

Do not improvise new taglines without CMO + CEO sign-off.

Email tone

Transactional emails from Beamix follow these rules:

  • Subject: specific, not clever. "Your GEO scan is ready" not "Exciting news about your AI presence"
  • Opening: no "Hi {first_name}, I hope this finds you well"
  • Closing: "— Beamix" not "Warmly, The Beamix Team" or "— your crew"
  • CTA button text: verb-first. "View your score" not "Click here to learn more about your score"
Subject: Your scan found 3 gaps (ChatGPT + Gemini)
Body:
Your latest GEO scan is ready.

ChatGPT: Not mentioned for "accounting in Haifa"
Gemini: Rank 4 for "tax help Tel Aviv" — up from rank 7
Perplexity: Not indexed

Ran queued 2 FAQ pages that address the ChatGPT and Perplexity gaps.

[Review your score →]

— Beamix

See also

  • copywriting — [[copywriting]]
  • marketing-psychology — [[marketing-psychology]]
  • humanizer — [[humanizer]]
  • seo-content-writer — [[seo-content-writer]]

Anti-patterns

  • Saying "— your crew" at the end of emails (use "— Beamix")
  • Naming agents on emails or PDFs (reserved for product UI only)
  • Adding AI disclosure labels on any generated content
  • Using English copy as the HE "translation" (full parity required)
  • Hedging language: "might help", "could potentially", "may improve" — be direct
  • Sentence-case headings on H1 (use Title Case for H1, Sentence case for H2–H6)