Training Your AI Agent
Learn how to effectively train your AI support agent with documents, URLs, Q&A pairs, and integrations for optimal performance.
Training belongs to the shared core and tenant knowledge layer
Knowledge training improves what the shared bot core can answer for a specific tenant. It does not replace explicit business actions or channel-specific adapters.
Document Upload
Upload PDFs, Word documents, text files, and more to train your agent.
Features
- Supported formats: PDF, DOC, DOCX, TXT, CSV, Markdown
- Maximum file size: 50MB per file
- Automatic text extraction and chunking
- OCR support for scanned documents
- Processing time: 1-5 minutes per document
Best Practices
- Organize documents by topic before uploading
- Use clear, well-structured documents
- Remove unnecessary formatting
- Break large documents into smaller sections
URL Crawling
Add URLs from your website or knowledge base for automatic content extraction.
Features
- Crawl single pages or entire websites
- Automatic link discovery and following
- Respect robots.txt and crawl delays
- Extract main content, ignore navigation
- Schedule periodic re-crawls for updates
Best Practices
- Start with your FAQ or help center pages
- Provide sitemap URLs for better crawling
- Exclude private pages and internal links
- Monitor crawl status and errors
Q&A Pairs
Manually add question-answer pairs for precise control over responses.
Features
- Add questions with multiple answer variations
- Import from CSV files
- Tag questions by category or topic
- Set confidence priorities
- Support for follow-up questions
Best Practices
- Cover your most common customer questions
- Write clear, concise answers
- Include multiple phrasings of same question
- Update regularly based on analytics
Integration Imports
Import knowledge from connected platforms like Notion, Zendesk, or Confluence.
Features
- Sync with Notion databases
- Import Zendesk help articles
- Pull from Confluence spaces
- Connect to Google Docs
- Automatic sync on schedule
Best Practices
- Choose relevant pages/articles only
- Set up automatic sync for updates
- Review imported content quality
- Remove outdated information
Training Best Practices
Follow these guidelines for optimal agent performance
Quality Over Quantity
Focus on accurate, well-written content rather than volume
Keep It Current
Regularly update training data to reflect changes
Test Responses
Use the test interface to verify agent accuracy
Monitor Performance
Review analytics to identify knowledge gaps
Organize by Topic
Structure training data into clear categories
Remove Duplicates
Avoid redundant information that confuses the agent
Users create, manage, and deploy bots
Sign up, upload documents, customize widgets, monitor analytics, and handle escalations.
Have questions? Our community is here to help. Learn more about our community
Active tenants
