Test Your Support BV
Test and preview your Support BV before deploying to customers
Test Your Support BV
Test your Support BV in a safe environment before deploying to customers.
Overview
The test interface lets you:
- Test conversations in real-time
- Verify AI responses and accuracy
- Check visual customization
- Simulate customer interactions
- Test all features (personality, training data, settings)
Accessing Test Interface
Navigate to Support BV
Select Support BV from your BrandVox dashboard.
Open Test Tab
Click the Test tab to access the testing interface.
Start Testing
The Support BV loads in an interactive window where you can start conversations immediately.
Testing Features
Standard View
Desktop-friendly view showing how Support BV appears on your website with full functionality.
Fullscreen Mode
Immersive testing experience filling your entire screen.
How to activate:
- Click Fullscreen button
- Test without distractions
- Press Esc to exit
What to Test
Conversation Flow
Test natural conversation progression:
- Simple questions - "What are your hours?"
- Complex inquiries - Multi-part questions
- Follow-up questions - Context retention
- Unclear questions - Clarification behavior
- Off-topic questions - Handling irrelevant topics
Response Accuracy
Verify correct, helpful information:
- ✅ Factually accurate information
- ✅ Current product details
- ✅ Correct policies (returns, shipping, terms)
- ✅ Proper handling of special cases
- ✅ Appropriate escalation when uncertain
If responses are inaccurate:
- Review training data completeness
- Check AI confidence threshold
- Add more examples for problematic topics
- Adjust personality settings
Brand Alignment
Ensure Support BV reflects your brand:
- Tone of voice - Professional, friendly, casual as intended
- Language - Uses your brand's terminology
- Personality - Matches enthusiasm/empathy level
- Visual design - Colors, fonts, and styling align
Visual Appearance
Test design elements:
- Colors - Primary/secondary colors display correctly
- Typography - Font loads, size is readable
- Layout - Proper bubble sizing and positioning
- Animations - Smooth opening/closing, typing indicator
Testing Best Practices
Create Test Scenarios
Develop structured testing plan:
- Welcome flow - Test greeting and starters
- Product questions - Ask about 5-10 products/services
- Support inquiries - Test common issues
- Edge cases - Unusual questions, typos
- Escalation - Request human assistance
- End conversation - Test goodbye and closure
Involve Your Team
Get multiple perspectives:
- Customer service team - Know common questions
- Marketing team - Ensure brand voice consistency
- Product team - Verify product information accuracy
- Random team members - Fresh perspective
Iterate and Retest
Testing is ongoing:
- Test thoroughly
- Identify issues
- Make adjustments
- Retest changes
- Repeat until satisfied
Common Testing Scenarios
New Setup (30-60 min)
- Focus: Basic functionality, appearance, brand voice, core knowledge
After Training Updates (15-30 min)
- Focus: New information accuracy, impact on existing responses
After Settings Changes (15-20 min)
- Focus: Impact of changes, unintended side effects
Before Major Launches (1-2 hours)
- Focus: Campaign info, high-traffic scenarios, team alignment
Interpreting Results
Good Signs (Ready to Deploy)
✅ Answers 80-90% of questions accurately ✅ Maintains brand voice consistently ✅ Handles follow-ups with context ✅ Escalates appropriately ✅ Loads quickly and performs smoothly ✅ Looks professional and on-brand
Warning Signs (Needs Refinement)
⚠️ Frequent "I don't know" responses ⚠️ Off-brand tone or language ⚠️ Slow response times ⚠️ Confusing or complex answers ⚠️ Visual glitches ⚠️ Inappropriate confidence levels
Red Flags (Do Not Deploy)
🚫 Factually incorrect information 🚫 Broken core functionality 🚫 Offensive or inappropriate responses 🚫 Major visual problems 🚫 Security or privacy concerns
Troubleshooting
After Testing
Once satisfied with performance:
- Document all changes made
- Get team sign-off
- Deploy to integration
- Monitor initial customer interactions
- Iterate based on real data