Knowledge Base Architecture
A well-structured knowledge base is the single biggest factor in AI support quality. This lesson covers how to architect one that scales.
The Three-Layer Model
Organize your knowledge base into three layers:
Layer 1: Quick Answers Short, direct answers to common questions. These should resolve in one message.
- "What are your business hours?"
- "How do I reset my password?"
- "What payment methods do you accept?"
Layer 2: How-To Guides Step-by-step instructions for completing tasks. Include screenshots where helpful.
- "How to set up two-factor authentication"
- "How to export your data"
- "How to upgrade your plan"
Layer 3: Deep Dives Comprehensive articles explaining concepts, troubleshooting complex issues, or covering policies in detail.
- "Understanding your billing cycle"
- "Troubleshooting API connection errors"
- "Data privacy and compliance overview"
Category Structure
Start with these universal categories and customize for your product:
- Getting Started (onboarding, first steps)
- Account & Billing (payments, plans, invoices)
- Features (one subcategory per major feature)
- Troubleshooting (organized by symptom)
- Policies (terms, privacy, SLAs)
- Integrations (one article per integration)
Naming Conventions
Consistent naming helps both AI retrieval and human browsing:
- Use question format for FAQs: "How do I...?" or "What is...?"
- Use action format for guides: "Setting up...", "Configuring..."
- Use descriptive format for reference: "Billing Policy", "API Rate Limits"
Scaling Considerations
- Plan for 50-200 articles in year one
- Assign ownership per category to prevent staleness
- Use tags for cross-category discoverability
- Archive outdated content rather than deleting it
Next up: How to write content that AI can understand and use effectively.