brand-voice

Defines the AE's personal communication voice, email formatting, vocabulary substitutions, and the 7-point voice check

SKILL.md

GitHub

Brand Voice — Communication Style Skill

Tier 3 — This entire file is regenerated during /setup Phase 4 based on the AE's communication preferences and writing samples.

Purpose

Defines the AE's personal communication voice for all written outputs. Contains voice rules, email formatting standards, vocabulary substitutions, banned phrases, and the 7-point voice check. Ensures every email, outreach, LinkedIn post, and follow-up sounds like the AE — not like generic AI.

When Referenced

  • /email — applies voice rules to contextual emails
  • /outreach — applies voice rules to prospecting emails
  • /draft-post — applies voice rules to LinkedIn posts
  • /closeout — applies voice rules to follow-up emails (Output 4)
  • /pov — applies voice tone to POV documents

Core Framework

Voice Rules

Communication style: {{COMMUNICATION_STYLE}}

Default voice (pre-/setup):

  • Direct and professional. No corporate fluff.
  • Short sentences. Active voice. Say what you mean.
  • Confident but not arrogant. Helpful but not subservient.
  • Match formality to the relationship — warm with champions, professional with economic buyers, concise with technical evaluators.

After /setup, these rules are rewritten based on the AE's actual communication preferences and writing samples.

Email Formatting

Rule Standard
Subject lines Lowercase unless proper nouns. Under 6 words. Specific, not clickbait.
Length Under 200 words. If it's longer, it should be a document, not an email.
Structure Context (1 sentence) → Purpose (1-2 sentences) → Ask (1 sentence)
Sign-off {{EMAIL_SIGN_OFF}}
Tone Varies by recipient — see Tone Matching below
CTA One clear call-to-action per email. Never more than one.

Tone Matching

Match communication tone to the recipient's role and relationship:

Audience Tone Example
Champion Warm, collaborative, insider language "Hey Sarah — quick update on the POC timeline..."
Economic Buyer Professional, business-outcome focused "Following up on our conversation about the projected ROI..."
Technical Evaluator Concise, specific, data-driven "Here are the integration specs you requested..."
New Contact Respectful, value-first, no assumptions "Reaching out because I noticed [specific observation]..."
Internal (manager) Structured, metrics-focused, brief "Pipeline update: 3 deals moved this week..."

Vocabulary Substitutions (10+ required)

After /setup, this section contains the AE's specific vocabulary preferences.

Default substitutions (pre-/setup):

Instead of Use
"Reach out" "Contact" or "get in touch"
"Circle back" "Follow up" or "revisit"
"Touch base" "Connect" or "check in"
"Synergy" [delete — don't replace, just remove]
"Leverage" "Use" or "apply"
"Bandwidth" "Capacity" or "time"
"Deep dive" "Detailed review" or "analysis"
"Ping" "Message" or "email"
"Loop in" "Include" or "add"
"Take offline" "Discuss separately"
"Ideate" "Brainstorm" or "think through"
"Socialize" "Share" or "discuss"

Banned Phrases

These phrases are NEVER used in any communication, regardless of context:

Default banned list (pre-/setup):

  • "Just checking in"
  • "Per my last email"
  • "Hope you're doing well"
  • "Wanted to follow up"
  • "At the end of the day"
  • "Low-hanging fruit"
  • "Move the needle"
  • "I hope this email finds you well"
  • "I wanted to reach out"
  • "As per our conversation"

After /setup, this list is customized with the AE's personal banned phrases and may add phrases they never want AI to generate.

7-Point Voice Check

Run this check on every communication output before presenting:

# Check Question
1 Authenticity Does this sound like {{AE_NAME}} wrote it, or like a robot?
2 Vocabulary Are all vocabulary substitutions applied? No banned phrases?
3 Length Is the email under 200 words? Is the LinkedIn post 150-300 words?
4 Structure One clear purpose? One clear CTA?
5 Tone match Is the tone appropriate for the recipient and relationship?
6 Sign-off Using {{EMAIL_SIGN_OFF}} (or appropriate LinkedIn close)?
7 Filler check Zero filler phrases? No "I hope," "just wanted to," "I'd love to"?

Passing score: 7/7 for emails and outreach. 5/7 acceptable for LinkedIn posts (length and sign-off rules differ).

LinkedIn Voice (Specific)

LinkedIn voice differs from email voice:

  • More conversational than email
  • First-person is fine (and encouraged for Personal posts)
  • Can be more opinionated
  • Can use incomplete sentences for emphasis
  • Emojis: sparingly, never more than 2 per post (unless AE prefers differently)
  • Hashtags: 3-5 at the end, never in the body

Personalization Notes

  • This entire file is Tier 3 — fully regenerated during /setup Phase 4 (Brand Voice)
  • /setup captures: communication samples, vocabulary preferences, banned phrases, sign-off, and style preferences
  • Voice rules are built from actual writing samples the AE provides during /setup
  • Vocabulary substitutions list grows over time as the AE identifies more words they don't use
  • The 7-point voice check is structurally universal but the specific checks reference personalized content

First-Use Calibration Storage (/outreach)

On the first run of /outreach, the user is asked for their typical email tone and optionally provides a sample email. The calibration response is stored here as supplemental voice data:

  • Outreach tone notes: [Populated on first /outreach run — natural email tone observations, sample email analysis]
  • Outreach-specific adjustments: [Any voice rules specific to prospecting emails vs. internal/existing-relationship emails]

This supplements the /setup voice profile. If /outreach runs before /setup, calibration data is stored here and merged into the full voice profile when /setup completes.