This guide is designed to help you fine-tune your AI agent so it communicates more closely with your business. While the AI already has a built-in context layer, the Knowledge Base and Agent Rules allow you to further align it with your specific preferences, voice, and goals.
1. How to Think About It
Knowledge Base = what your agent knows
Agent Rules = how your agent behaves
The Knowledge Base provides context and content, while Agent Rules shape how that content is delivered.
2. Knowledge Base (The Brain)
This is where you provide the materials your agent uses when communicating on your behalf.
What it impacts:
LinkedIn outreach (Outbound Agent)
Email follow-ups (Inbound Agent)
Chatbot + Web Pixel responses
Tone, phrasing, and positioning
Tip: If your agent may need to give customers a phone number, office address, or other contact detail, make sure that information is included in the document you upload here. The agent can only share contact details that are actually in its Knowledge Base
How to add content:
Go to Knowledge Base
Click Upload Files
Upload your file
Select which agent(s) it applies to
Click Upload
What to include:
Product / service documents
Messaging examples (LinkedIn, email, scripts)
FAQs and objections
Brand voice guidelines
Competitor context
Internal positioning
What you don’t offer
Pro tip: Include what you DON’T offer. This prevents hallucinations and improves qualification accuracy.
3. Agent Rules (Behavior + Guardrails)
This is where you guide how the agent shows up in conversations.
What it controls:
Tone (casual, professional, conversational)
Selling style (soft vs direct)
Language preferences
Behaviors to follow or avoid
What to define:
Tone
Example: “Conversational, thoughtful, not overly sales-driven”
Selling approach
Example: “Focus on providing value and building relationships rather than pushing for a sale”
Do / Don’t guidelines
Avoid buzzwords, remove robotic phrasing, and eliminate unwanted habits like overused punctuation.
Preferred terminology
Define words to use consistently and terms to avoid.
Be specific in your guidance, use this section to correct recurring AI habits, and define selling behavior here rather than relying on your documents to control it.
4. How They Work Together
The Knowledge Base provides context, while Agent Rules shape how that context is expressed.
If something feels off:
Content issue → update the Knowledge Base
Tone or behavior issue → adjust Agent Rules