Skills

Foundation Brand Voice

elevate-foundation-brand-voice

FoundationElevate

Define the FOUNDATION Brand Voice — three to five actionable adjectives plus do/don't guidance — for cell C2 of the Company Context Grid. Use when fixing how the brand sounds.

Price
Included no per-call charge
Evals
0 none shipped yet
Assertions
0 nothing to check
References
0 self-contained
Length
661 words · 3 min read
Sections
3 in the instructions

Evals

0 cases

No evals ship with this skill yet. Edit it to add test cases, then Evaluate scores it against them.

Details

Family
Foundation
License
MIT
Trigger
semantic
Source
ai/skills/elevate-foundation-brand-voice/SKILL.md
Eval bank
ai/skills/elevate-foundation-brand-voice/evals/evals.json
Price
Included — no per-call charge
References
None — self-contained

Instructions

Define the Brand Voice

Objective. Fix the brand's voice as three to five specific, actionable adjectives — each with do/don't examples and channel notes — so every person and AI tool downstream sounds like the same brand: cell C2 of the Company Context Grid.

Inputs this skill needs

  • [Company Context] — the Foundation's company slot: [Brand Archetype & Pillars], [Mission & Vision], any existing communications samples (About page, product descriptions, emails, social posts).
  • No upstream skills (Wave 1).

ROCKET prompt

ROLE: You are a brand strategist and communication designer who turns a brand's personality into language other people and AI tools can reproduce. You name a voice precisely enough that it changes how a sentence is written.

OBJECTIVE: Produce the Brand Voice definition for cell C2 of the Company Context Grid — three to five voice adjectives that are specific and actionable, each with positive and negative examples and a note on how it bends per channel, expressing the archetype defined in B3.

CONTEXT: Work from the supplied [Company Context], especially the [Brand Archetype & Pillars] and any sample communications. This is the C2 cell the Company chapter calls the most immediately practical in the Grid: the difference between content that sounds like the brand and content that sounds like its category. The archetype (B3) must be legible in the voice — a Sage sounds like a Sage, not a Rebel. Do not duplicate the customer-facing prose rules; this asset is consumed by the elevate-voice skill and every later step as the brand's standing voice context.

KEY INSTRUCTIONS:

  1. Choose three to five adjectives that are specific enough to be actionable. Reject the decorative defaults: not "professional", not "friendly". Favour adjectives that rule things out as clearly as they rule things in (warm but direct; specific over vague; confident without posturing; plain-spoken over jargon).
  2. Apply the chapter's test to every adjective: does it actually change how a sentence would be written? If it would not change the writing, cut it as decorative.
  3. For each adjective give: a one-line behavioural definition, two "sounds like" examples, and two "never sounds like" examples — so the voice is teachable and its boundaries clear.
  4. If sample communications are supplied, calibrate to the voice the brand actually has, not the one it wishes for — name any gap between the two and recommend which to write toward.
  5. Note, in one line per adjective, how it modulates across channels (the voice is constant; its dress changes by surface) — the brief that Market's Row C channel work will draw on.
  6. End with one synthesis line: the voice in a single sentence a writer can hold in mind.

EXAMPLES (illustrative shapes, not branded):

  • "Direct — opens with the point, cuts the preamble. Sounds like: 'This lasts ten years. Here's why.' Never sounds like: 'We're so excited to share a little something with you today...'"
  • "Precise without being cold — names the mechanism, keeps the warmth. Sounds like: 'Steeped 36 hours, so it's never bitter.' Never sounds like: 'Our revolutionary process unlocks next-level flavour.'"

TONE & FORMAT: Analytical and practical — the brand-strategy register. British English. The asset itself is internal voice documentation; any sample copy inside it must pass elevate-voice (no hype register, no exclamation marks). Structure exactly as the Output contract.

Output contract

Write to companies/<slug>/playbook/0-foundation/brand-voice.md:

  • # Brand Voice (Company Context Grid — C2)
  • ## Voice adjectives — a numbered list of 3–5 adjectives. Under each: Definition (one line), Sounds like (two examples), Never sounds like (two examples), Across channels (one line).
  • ## Voice as actually written (include only if samples were supplied) — one short paragraph naming any gap between current and intended voice, and which to write toward.
  • ## The voice in one sentence — a single synthesis sentence.

Total length 250–400 words. Specific enough that elevate-voice and every downstream step can adopt the voice without further explanation.