Skip to main contentThe Agent Playground is your testing environment where you can interact with your agents before deploying them to your website. It’s the perfect place to test, refine, and improve your agents to ensure they provide the best possible experience for your users.
What Is the Agent Playground?
The Agent Playground is a safe testing environment where you can:
- Chat with your agents in real-time
- Test different scenarios and questions
- Verify agent responses and behavior
- Refine agent instructions based on testing
- Ensure agents work correctly before going live
Recommended Workflow
For the best results, we recommend this testing sequence:
- Run simulations - Let your agents explore and learn your application interface first
- Test in the playground - Verify that simulation learning improved agent performance
- Refine and iterate - Make adjustments based on testing results
- Deploy to production - Launch your well-tested agents to help real users
This approach ensures your agents have both knowledge and real interface understanding before they start helping users.
Why Use the Playground?
Before deploying your agents, the playground helps you:
- Test agent responses to ensure they’re accurate and helpful
- Identify issues and areas for improvement
- Refine agent behavior based on real interactions
- Build confidence that your agents will work well with real users
- Save time and resources by catching problems early
How the Playground Works
The Testing Process
- Select your connection - Choose which website or app to test with
- Select your agent - Choose which agent to test
- Start chatting - Begin testing with sample questions
- Test different scenarios - Try various types of questions and situations
- Review responses - Check that agent responses are helpful and accurate
- Refine as needed - Update agent instructions or knowledge based on testing
Playground Features
Real-Time Chat Interface
Interactive testing:
- Chat with your agents in real-time
- See responses immediately
- Test different conversation flows
- Experience the user journey
Conversation history:
- View previous messages
- Track conversation context
- Review agent responses
- Analyze conversation patterns
Connection and Agent Selection
Flexible testing:
- Test with different connections
- Switch between different agents
- Compare agent performance
- Test specific configurations
Easy switching:
- Change connections quickly
- Switch agents without losing context
- Test different combinations
- Compare results easily
Response Analysis
Quality assessment:
- Evaluate response accuracy
- Check response helpfulness
- Assess response completeness
- Review response tone and style
Performance tracking:
- Monitor response times
- Track conversation success
- Identify improvement areas
- Measure agent effectiveness
What to Test in the Playground
Basic Functionality
Test fundamental capabilities:
- “Hello, can you help me?”
- “What does your company do?”
- “How can I contact support?”
- “What are your business hours?”
Verify basic responses:
- Agent responds appropriately
- Responses are helpful and accurate
- Tone matches your brand
- Information is correct
Knowledge Base Testing
Test knowledge usage:
- “What products do you offer?”
- “How do I get started?”
- “What are your policies?”
- “Can you help me with [specific topic]?”
Verify knowledge integration:
- Agent uses your knowledge base
- Responses reference your content
- Information is accurate and current
- Knowledge is relevant to questions
Interaction Mode Testing
Test Tell Mode (Answering Questions):
- “What are your business hours?”
- “How do I reset my password?”
- “What features does this have?”
- “Can you explain how this works?”
Test Show Mode (Demonstrating Tasks):
- “Show me how to create an account”
- “Show me how to use this feature”
- “Show me how to navigate your website”
- “Show me how to set up the product”
Test Do Mode (Performing Actions):
- “Help me fill out this form”
- “Guide me through the checkout process”
- “Automate this task for me”
- “Complete this process for me”
Edge Case Testing
Test challenging scenarios:
- “I don’t understand what you’re saying”
- “Can you help me with something not related to your business?”
- “What if I’m not satisfied with your service?”
- “I’m having trouble with [specific issue]”
Verify error handling:
- Agent handles unclear questions
- Agent stays on topic appropriately
- Agent escalates when needed
- Agent provides helpful alternatives
Playground Best Practices
Systematic Testing
Test systematically:
- Start with basic questions
- Progress to more complex scenarios
- Test different interaction modes
- Try edge cases and challenging questions
Document your testing:
- Keep notes on what works well
- Record issues and problems
- Track improvement areas
- Document successful test cases
Iterative Improvement
Test, refine, and repeat:
- Test your agent thoroughly
- Identify areas for improvement
- Update agent instructions or knowledge
- Test again to verify improvements
Continuous improvement:
- Regular testing and refinement
- Monitor agent performance
- Update based on real user feedback
- Continuously optimize
Real-World Scenarios
Test realistic scenarios:
- Use questions real users might ask
- Test common use cases
- Try different user personas
- Simulate real conversations
Prepare for real users:
- Test with various question types
- Verify agent can handle different situations
- Ensure responses are helpful and accurate
- Build confidence in agent performance
Common Testing Scenarios
Customer Support Testing
Product questions:
- “What are the features of this product?”
- “How do I choose the right size?”
- “What’s included with this purchase?”
- “How do I set up this product?”
Order support:
- “How do I track my order?”
- “Can I change my shipping address?”
- “What’s your return policy?”
- “How do I cancel my order?”
Account management:
- “How do I reset my password?”
- “Can I update my account information?”
- “How do I change my email address?”
- “Can I delete my account?”
Sales and Marketing Testing
Service information:
- “What services do you offer?”
- “How much does this cost?”
- “What are your packages?”
- “Do you offer discounts?”
Feature explanations:
- “What makes you different from competitors?”
- “How does this feature work?”
- “What are the benefits of this service?”
- “Can you explain your pricing?”
Lead generation:
- “How do I get started?”
- “Can you send me more information?”
- “Do you offer free trials?”
- “How can I contact sales?”
Technical Support Testing
Troubleshooting:
- “Why isn’t this working?”
- “How do I fix this error?”
- “What should I do if this happens?”
- “Can you help me troubleshoot?”
Setup and configuration:
- “How do I set this up?”
- “What do I need to configure?”
- “How do I connect this to my system?”
- “What are the requirements?”
Integration help:
- “How do I integrate this with my other tools?”
- “What APIs are available?”
- “How do I set up the connection?”
- “Can you help me with the integration?”
Playground Limitations
Testing Environment
Not production:
- Playground is for testing only
- Not connected to real user data
- Limited to test scenarios
- May not reflect real-world performance
Limited context:
- No real user history
- No actual website integration
- Limited to test conversations
- May not show full capabilities
Response times:
- Playground may have different performance
- Real-world usage may vary
- Network conditions may differ
- Load may affect performance
Scalability:
- Playground doesn’t test scalability
- Real users may create different load
- Performance may vary under load
- Consider production requirements
Next Steps
Now that you understand the playground, you’re ready to:
- Test your agents - Learn how to test your agents effectively
- Use the chat interface - Master the playground interface
- Refine your agents - Improve your agents based on testing
- Create integrations - Deploy your tested agents to your website
The playground is your key to creating successful AI agents. Use it thoroughly to test, refine, and perfect your agents before they go live.
Getting Help
If you need assistance with the playground:
- Check the troubleshooting guide for common issues
- Review testing best practices for effective testing
- Test systematically to identify issues
- Contact support for additional help
You’re now ready to test your agents effectively in the playground! 🚀 
