Setup Guide & Best Practices
A comprehensive guide to setting up your AI customer service assistant for optimal performance.
Setup Database
Initialize your knowledge base with business info
Upload Knowledge
Add documents and FAQs
Connect Channels
Link Gmail, Telegram, etc.
Configure & Test
Fine-tune AI behavior
Step 1: Create Your Business Information Document
Important: Before anything else, create a concise text document (around 100 words) describing your business. This gives your AI the essential context it needs from the start.
Your document should include:
- Company name and what you do
- Your website URL
- Main products or services
- Key contact information
- Business hours or availability
Example (90 words):
"Acme Corp is an e-commerce retailer specializing in sustainable home goods. Visit us at acmecorp.com. We offer eco-friendly kitchenware, cleaning supplies, and home decor with free shipping on orders over $50. Returns accepted within 30 days. Prices range from $10-$200. Customer service available Mon-Fri 9AM-6PM EST at support@acmecorp.com or 555-0123. We ship to US and Canada. Our mission is making sustainable living accessible and affordable for every household."
Step 2: Upload to Knowledge Base
Once you've created your business information document, upload it to your knowledge base. This becomes the foundation for all AI responses.
Supported formats:
- • Text files (.txt) - Simple and effective
- • PDF documents (.pdf) - Professional format
- • Word documents (.docx) - Easy to edit
Additional Documents to Upload
- Frequently asked questions (FAQs)
- Product catalogs and descriptions
- Return and shipping policies
- Pricing information
- Terms of service
Best Practices
- • Keep information current and accurate
- • Use clear, customer-friendly language
- • Include specific examples and scenarios
- • Organize by topic or category
- • Update regularly as business changes
Email (Gmail)Recommended
- • Most common customer support channel
- • Automatic email threading
- • Professional communication
Note: You can connect multiple channels. Each channel will have access to the same knowledge base and AI configuration.
Tone & Style
- • Set professional vs. casual tone
- • Maintain brand voice consistency
- • Choose language preferences
- • Define ideal response length
System Prompts
- • Define AI role and personality
- • Set response guidelines
- • Add company-specific instructions
- • Configure escalation triggers
Behavior Settings
- • Set escalation rules
- • Configure response time expectations
- • Define follow-up procedures
- • Establish error handling protocols
Advanced Options
- • Channel-specific behavior
- • Notification preferences
- • Data collection settings
- • Privacy controls
Initial Testing
- • Send test messages to each channel
- • Verify response accuracy
- • Check escalation triggers
- • Test different question types
- • Validate knowledge base usage
Quality Checks
- • Review response tone and style
- • Check for brand consistency
- • Verify accuracy of information
- • Test edge cases and unusual queries
Ongoing Monitoring
- • Review conversation logs daily
- • Monitor customer satisfaction scores
- • Track response times
- • Update knowledge base regularly
Analytics & Insights
- • Track conversation volume
- • Measure resolution rates
- • Identify common questions
- • Review performance metrics
Knowledge Management
- • Start with essential business info (100 words)
- • Keep all information current and accurate
- • Use clear, customer-friendly language
- • Include specific examples and scenarios
- • Organize content by topic and priority
AI Configuration
- • Provide clear role definitions
- • Set specific response guidelines
- • Define escalation procedures clearly
- • Test thoroughly before going live
- • Review and update settings regularly
Customer Experience
- • Maintain consistent brand voice
- • Provide helpful, accurate responses
- • Escalate complex issues appropriately
- • Follow up on unresolved matters
- • Monitor satisfaction scores
Continuous Improvement
- • Regular performance reviews
- • Analyze customer feedback
- • Update knowledge base monthly
- • Track key metrics and KPIs
- • Iterate based on data insights
Start small: Begin with your core business information document, then gradually add more detailed documents as you identify customer needs.
Test with real questions: Use actual customer questions from past interactions to test your AI's responses.
Update frequently: Set a reminder to review and update your knowledge base monthly, especially after product launches or policy changes.
Monitor the first week: Check conversations daily during the first week to catch any issues early and refine your setup.