Overview
Creating a knowledge base is your first step toward giving your AI agent expert-level knowledge of your business. This guide walks you through the complete process, from creating the knowledge base to organizing content and following best practices.Step 1: Create a Knowledge Base
Navigate to Knowledge Bases
- Log into your itellicoAI dashboard
- Click Knowledge Bases in the left navigation menu
- Click the + Create Knowledge Base button
Fill in Basic Information


Enter Knowledge Base Name
Choose a clear, descriptive name that reflects the content category.Naming rules:
- Letters, numbers, spaces, hyphens, and underscores only
- No special characters (@, #, $, etc.)
- Maximum 100 characters
- “Customer Support FAQ”
- “Product Documentation”
- “Company Policies”
- “Technical Troubleshooting”
- “KB_001” (not descriptive)
- “Misc Docs” (too vague)
- “Everything!!!” (unprofessional)
Add Description (Optional)
Provide context about what this knowledge base contains and when to use it.Example descriptions:
- “Customer-facing support documentation including FAQs, policies, and troubleshooting guides”
- “Internal product specifications, feature documentation, and technical details”
- “Company HR policies, compliance documents, and internal procedures”
Step 2: Organize with Folders
Folders organize your knowledge items into logical groups.A default folder is automatically created when you create a new knowledge base. You can rename it, add more folders, or start adding items right away.
Add Additional Folders (Optional)
- Open your knowledge base
- Click the Create Folder button
- Enter a folder name
- Add an optional description
- Click Create
Folder Naming Best Practices
Clear & Specific
- “Billing & Payments”
- “Return Policies”
- “Product Specifications”
- “Technical Issues”
Vague or Generic
- “Stuff”
- “Folder 1”
- “Docs”
- “Misc”
How Many Folders?
Small Knowledge Base (< 50 items)
Small Knowledge Base (< 50 items)
3-5 folders is usually sufficientExample structure:
- Customer Questions
- Company Policies
- Product Information
Medium Knowledge Base (50-200 items)
Medium Knowledge Base (50-200 items)
5-10 folders provides good organizationExample structure:
- Billing & Payments
- Technical Support
- Product Features
- Account Management
- Shipping & Returns
- General FAQ
Large Knowledge Base (200+ items)
Large Knowledge Base (200+ items)
10-15 folders keeps things manageableConsider splitting into multiple knowledge bases if you exceed 15 folders.Example structure:
- Account Setup
- Billing Issues
- Payment Methods
- Refunds & Credits
- Product Categories A-F
- Product Categories G-M
- Product Categories N-Z
- Technical Troubleshooting
- Integration Guides
- API Documentation
- Returns & Exchanges
- Shipping Options
- Order Tracking
Step 3: Add Knowledge Items
Knowledge items are the actual content - the documents, FAQs, policies, and information you want your agent to know.

Text Entry - Write content directly
Text Entry - Write content directly
Best for: FAQs, policies, procedures, quick reference informationDirect text entry using the built-in editor. Perfect for content you’re creating specifically for your knowledge base.
How to Add a Text Item
Enter a descriptive title
Good titles:
- “Return Policy for Digital Products”
- “Password Reset Instructions”
- “Pricing Tiers Explained”
- “Doc 1”
- “Policy”
- “Info”
Write or paste your content
Enter the information you want your agent to know. Be clear and comprehensive.Example:
File Upload - Upload existing documents
File Upload - Upload existing documents
Best for: User manuals, product specs, legal documents, training materialsUpload documents in various formats up to 10MB. The system extracts and processes the text content.Supported formats: PDF, Word (.doc, .docx), Text (.txt, .log), Markdown (.md), CSV/TSV, JSON, YAML, XML
How to Add a File
Upload your file
Click Upload and select your document.Requirements:
- Supported formats: PDF, Word (.doc, .docx), Text (.txt, .log), Markdown (.md), CSV/TSV, JSON, YAML, XML
- Maximum size: 10MB
- Text-based documents and scanned images (advanced parsing handles most scans)
URL Scraping - Pull content from web pages
URL Scraping - Pull content from web pages
Best for: Documentation pages, blog posts, help articles, external resourcesScrape content from a single web page (not entire websites).
How to Add a URL
URL scraping works best with simple, text-based web pages. Complex JavaScript applications or pages requiring authentication may not scrape successfully.
Understanding Processing Statuses
After adding a knowledge item, it goes through two separate processing pipelines. You’ll see both statuses in the interface:Content Processing Status
This tracks whether the content was successfully extracted from your source.PENDING
PENDING
Item created and queued for processing. Usually takes just a few seconds.
PROCESSING
PROCESSING
Content is being processed right now:
- Files: Parsing documents using our advanced AI document parsers
- URLs: Fetching and scraping the web page
- Text: Skips this (goes directly to COMPLETED)
COMPLETED
COMPLETED
Content successfully extracted and stored. Vector indexing will begin automatically.
FAILED
FAILED
Content extraction failed. Common causes:
- File is corrupted or password-protected
- URL requires authentication or uses complex JavaScript
- File exceeds 10MB limit
- Try uploading the content as TEXT instead
- Remove password protection from the file
- Split large files into smaller documents
Vector Indexing Status
After content is extracted, it must be indexed for RAG (Retrieval-Augmented Generation). This allows your agent to find relevant knowledge based on meaning.PENDING
PENDING
Waiting for vector indexing to start. Content processing completed successfully.
INDEXING
INDEXING
Creating vector embeddings for RAG:
- Content is split into chunks
- AI embeddings are generated
- Vectors are stored in the database
INDEXED
INDEXED
Item is fully ready! The agent can now find and use this knowledge via semantic search.
FAILED
FAILED
Vector indexing failed. The content is extracted but won’t appear in RAG search results.Quick fixes:
- Click the Reindex button on the knowledge item to retry indexing
- If reindexing fails, try deleting and re-creating the item
- Ensure the content isn’t too short (needs at least a few sentences)
Both statuses must succeed for full functionality:
- Content Status: COMPLETED
- Vector Status: INDEXED
Content Quality Guidelines
Write Clear, Comprehensive Content
Be specific and detailed
Be specific and detailed
Don’t assume knowledge. Explain clearly.Instead of:
“Customers can return items within our standard window.”Write:
“Customers can return items within 30 days of purchase. Items must be unused and in original packaging with all tags attached.”
Use plain language
Use plain language
Avoid jargon unless it’s industry-standard and your audience knows it.Instead of:
“Utilize the authentication endpoint to generate a bearer token.”Write:
“To authenticate, call the /auth endpoint. You’ll receive an access token to include in subsequent requests.”
Include examples
Include examples
Examples make abstract concepts concrete.Example:
“Our pricing tiers include:
- Basic: $29/month - Up to 1,000 contacts
- Pro: $99/month - Up to 10,000 contacts
- Enterprise: Custom pricing - Unlimited contacts”
Structure with headings
Structure with headings
Use headings and bullet points for scannable content.Good structure:
Keep content focused
Keep content focused
Each item should cover one topic thoroughly. Split large topics into multiple items.Instead of:
One 20-page “Complete Company Handbook”Create:
- Return Policy
- Shipping Options
- Payment Methods
- Product Warranty
- Privacy Policy
- Contact Information
Link Knowledge Base to Agents
Once your knowledge base has content, you’ll need to link it to your agents so they can access the information.Assign Knowledge to Agents
Learn how to connect your knowledge base to agents and choose the right access method (Context vs RAG).