> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getwhys.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Brand Voice

> Encode how your company writes. Used by Message Testing to flag drafts that drift off-brand.

Brand Voice is the structured description of how your company sounds in writing. Each characteristic has a name, description, do's, and don'ts. Message Testing uses Brand Voice to evaluate whether a draft sounds like it came from your company — and to flag where it drifts.

## Why structured Brand Voice matters

Voice is the thing AI tools get wrong the fastest. Trained on the internet, they default to generic. Even when you prompt them with "write in our voice," they tend to produce competent B2B prose that could have come from any of your competitors.

Encoding your Brand Voice in GetWhys turns it into a check, not a suggestion. Every Message Test evaluates content against the actual characteristics you've defined and flags the lines that violate the do's and don'ts.

## Structure

A Brand Voice is made up of several **Characteristics**. Each one has:

* **Name** — one to five words capturing the characteristic (e.g., "Direct", "Warm but expert")
* **Description** — what the characteristic means in practice
* **Do's** — examples of language and patterns that match
* **Don'ts** — examples of language and patterns to avoid

Most brand voices have three to five characteristics. More isn't always better — a few sharp characteristics with concrete examples produce better feedback than a long list of fuzzy ones.

## Creating a Brand Voice

Go to **Customization** in the left navigation, click **Brand Voice**, then **Add Characteristic**. Add a name, description, do's, and don'ts. Save. Repeat for each characteristic until your voice is fully described.

## Don't have a brand voice document?

Use one of the Chat templates:

<CardGroup cols={2}>
  <Card title="Existing Brand Voice" icon="file-lines">
    If you already have brand guidelines in some other format, this template translates them into the structured GetWhys format. Open Chat → **View all templates** → expand **Brand Voice** → pick **Existing Brand Voice**. Paste your guidelines and run.
  </Card>

  <Card title="Infer Brand Voice" icon="wand-magic-sparkles">
    If you don't have formal brand guidelines, this template will infer your voice from sample content. Open Chat → **View all templates** → expand **Brand Voice** → pick **Infer Brand Voice**. Paste your homepage copy, a couple of blog posts, and some social posts.
  </Card>
</CardGroup>

## Drafts

Save a characteristic as a draft while you're still iterating. Drafts don't affect Message Testing until you publish them.

## How Brand Voice is used

When you run a [Message Test](/features/message-testing), GetWhys evaluates your content against the active Brand Voice characteristics and flags lines that violate the do's and don'ts. You see which characteristics match, where it drifts, and quoted examples from the content itself.

Through MCP, Brand Voice also gets pulled into any AI tool that calls into GetWhys, so content generated in Claude, Cursor, or a custom agent starts from the right voice. See [MCP integration](/integrations/mcp).

## Tip

<Tip>
  If Message Testing feedback feels generic, your Brand Voice setup is usually the place to look first. Short, vague descriptions with thin do/don't lists produce thin feedback. Detailed, opinionated descriptions with specific examples produce specific feedback.
</Tip>