Back to Guides
Knowledge Base Best Practices
Learn how to structure your knowledge base for optimal AI performance and accurate responses.
Why Knowledge Quality Matters
Your chatbot is only as good as the knowledge you provide it. Well-structured, comprehensive knowledge leads to:
- More accurate and relevant responses
- Better customer satisfaction
- Reduced need for human intervention
- Increased lead conversion rates
Content Structure Guidelines
1. Use Clear, Descriptive Titles
Good Examples:
- "Pricing Plans - Monthly and Annual Options"
- "How to Reset Your Password"
- "Shipping Times for International Orders"
Bad Examples:
- "Info" (too vague)
- "Page 1" (not descriptive)
- "Untitled Document" (meaningless)
2. Optimal Content Length
Balance detail with conciseness:
- Minimum: 100 words (too short lacks context)
- Optimal: 300-1000 words per source
- Maximum: 2000 words (longer content should be split)
3. Use Categories Effectively
Organize content into logical categories:
Product Categories
- • Features
- • Pricing
- • Technical Specs
Support Categories
- • Getting Started
- • Troubleshooting
- • FAQs
Best Practices
Keep Content Updated
Review and update your knowledge base monthly. Remove outdated information and add new products/services.
Write in Natural Language
Write as if explaining to a customer. Avoid jargon unless necessary, and define technical terms.
Include Common Questions
Add FAQ-style content for common customer questions. This improves response accuracy.
Test Your Bot Regularly
Ask your chatbot common questions and evaluate responses. Adjust knowledge base based on results.
What to Avoid
- ❌Duplicate Content: Don't repeat the same information in multiple sources
- ❌Overly Technical Language: Unless your audience is technical, keep it simple
- ❌Empty Placeholders: Remove sources with "Coming Soon" or placeholder text
- ❌Too Many Sources: Quality over quantity - 20-30 well-written sources is better than 100 poor ones