AI Agents & Chatbots
Stop the assistant from sounding generic.
The Agent Prompt is the system prompt that teaches AI how to represent the brand. Paste it into your chatbot, support tool, or assistant so the response style stops drifting.
Support bots that stop sounding fake
What breaks
Your support bot sounds like a cheerful assistant. Your brand is calm and direct.
What to do
Use the Agent Prompt as the system prompt. That is where the voice has to lock.
What the prompt carries
- Brand archetype and response posture
- Voice rules: direct, calm, never a cheerleader
- Banned words list
- Example responses in your voice
- Constraints such as no urgency and no exclamation points
User: "Where's my invoice?"
Bot: "Great question! I'd be happy to help you with that! š"
User: "Where's my invoice?"
Bot: "Check your email for the receipt. It arrives within 5 minutes of purchase. If it's not there, I'll resend it now."
In-app assistants that match the product
What breaks
You built an AI assistant into your product. It sounds generic.
What to do
Feed it the Agent Prompt before you ship it to users.
Example
Product: A project management tool for developers
Archetype: Creator + Ruler
Result: the AI sounds like the product team, not a motivational coach.
Documentation helpers with actual authority
What breaks
Your docs bot explains things too casually. Your brand is technical and precise.
What to do
Apply the Voice Rules so the documentation stays precise instead of getting chatty.
User: "How do I authenticate?"
"No worries! Just head over to settings and you'll find the API key section. Easy peasy!"
User: "How do I authenticate?"
"Go to Settings ā API Keys. Copy the key. Add it to your request header as 'Authorization: Bearer [key]'."
Ready to stop sounding like every other bot?
Build the Brand Schema. Then export the Agent Prompt.
See the demo