Brand Voice
Ensure every AI-generated personalization sounds like your brand. Without Brand Voice, the AI might write great copy that doesn't match your tone. Brand Voice settings constrain the AI so it always stays on-brand.
Where to configure
Dashboard > Brand — the Brand page for your site.
1. Tone of Voice
A free-text field where you describe your desired tone in your own words (up to 200 characters). This affects word choice, sentence structure, and formality level.
Example Tones
- “Professional but friendly, confident without being aggressive”
- “Casual and conversational, like talking to a colleague”
- “Direct and technical, for a developer audience”
2. Personalization Intensity
Controls how aggressively the AI rewrites content.
Minimal
Minor tweaks to match intent. Safe, subtle changes that preserve your original copy almost entirely.
Moderate
Meaningful rewrites that preserve structure. Sentences may be rephrased but the overall flow stays the same.
Full
Bold rewrites that may restructure sentences entirely. Best for landing pages where conversion is the top priority. This is the default setting.
3. Brand Guidelines
Free-text instructions the AI follows on every generation. Be as specific as possible — vague guidelines produce vague results.
Example Guidelines
- “Always use action verbs in headlines”
- “Never mention competitor names”
- “Keep sentences under 15 words”
- “Use 'you' and 'your' instead of 'customers' or 'users'”
- “Our product name is always capitalized: GetIntent”
4. Preferred Keywords
Words and phrases the AI should try to include when they fit naturally. These are suggestions, not requirements — the AI won't force them into every sentence.
How it Works
When the AI generates personalized content, it receives your brand profile as context alongside the visitor's intent. The AI is instructed to:
- 1Match your tone setting
- 2Follow your brand guidelines exactly
- 3Include preferred keywords where natural
- 4Never use restricted keywords
- 5Fall back to your original content if it can't generate something on-brand
Learning from Feedback
The Transformation History page (Dashboard > Pageviews) shows every personalization the AI generated. You can approve or reject each one:
Approved
Patterns become examples the AI learns from and repeats in similar contexts.
Rejected
Patterns are avoided in future generations, helping the AI improve over time.
Tips
- Be specific in guidelines — “Use action verbs” is better than “Write well”
- Start with medium intensity and adjust based on results
- Review the first 20–30 transformations and give feedback to train the AI
- Different sites can have different brand voices — useful for agencies managing multiple clients