A brand voice template for AI that describes decisions, not adjectives.
Most AI brand-voice prompts begin with “professional, friendly and authentic.” So do thousands of other brands.
A useful voice guide shows the model what choices the brand tends to make: how direct it is, what it refuses to exaggerate, which words feel natural and what real examples sound like.
Adjectives describe a vibe. They do not resolve a sentence.
Two brands can both be warm and confident while writing completely differently. One uses short practical sentences; another tells stories. One says 'buy now'; another never uses urgency.
AI needs examples and boundaries that change actual writing decisions.
The voice sheet to keep beside the model.
BRAND VOICE FOR AI
WHO WE ARE
One sentence describing the business and who it helps.
VOICE PRINCIPLE 1:
What this means in writing:
Example:
VOICE PRINCIPLE 2:
What this means in writing:
Example:
VOICE PRINCIPLE 3:
What this means in writing:
Example:
WORDS / PHRASES WE NATURALLY USE:
-
-
LANGUAGE WE AVOID:
-
-
CLAIMS BOUNDARY:
What we never imply without evidence.
REAL EXAMPLE WE LIKE:
Paste 2–5 sentences that genuinely sound like us.
VISUAL CHARACTER:
How should imagery, typography and colour generally feel without forcing one template?Example: a straightforward local service brand.
WHO WE ARE: We help busy households keep their homes clean without making the process complicated.
VOICE 1: Direct.
MEANS: Say what the service includes before adding persuasive language.
EXAMPLE: 'A standard clean covers the rooms you use every day.'
VOICE 2: Calm.
MEANS: No panic, guilt or fake urgency.
EXAMPLE: 'Book when the house needs it. Weekly plans are optional.'
VOICE 3: Practical.
MEANS: Prefer concrete examples over lifestyle slogans.
AVOID: 'Transform your life', 'luxury experience', 'limited slots!!!'
CLAIMS: Never promise a perfectly germ-free home.
VISUAL: Natural light, real domestic spaces, restrained typography; no glossy stock-family look.Make the template earn its place.
Use examples that survived real use.
A sentence from a website, email or post the team already likes teaches more than a list of personality adjectives.
Add negative boundaries.
Knowing what the brand refuses to sound like is often the fastest way to avoid generic AI habits.
Keep facts separate.
Voice guidance should not become the source of product claims. Store business facts and claims boundaries explicitly.
A template dies when every possible field becomes mandatory.
- Three strong voice principles are usually enough.
- Do not list every word the brand has ever used.
- Avoid contradictory instructions such as 'minimal but highly detailed.'
- Do not turn visual guidance into one compulsory layout.
How many examples should I give AI for brand voice?
A few representative examples are usually more useful than a large uncurated archive. Pick examples that show the decisions you want repeated.
Should I include banned words?
Only when they reflect a real brand boundary. A huge banned-word list often creates awkward writing rather than stronger voice.
Can Woven store this context?
Yes. Voice guidance, business facts, colours, preferences and recent work can form part of Woven's Brand Knowledge.