3Chat.AI Sandbox Testing Function Introduction
3Chat.AI is a dynamic AI customer service agent designed specifically for Chinese sellers.
It breaks through the traditional passive response logic of customer service:
-
Pre-sales phase: Executes intelligent marketing SOP, combining industry-specific sales scripts and multi-turn conversational pacing to proactively guide customers to leave contact information and convert.
-
Post-sales phase: Acts as a business expert, performing business operations based on business goals to help improve customer retention and repurchase rates.
Before going live, sellers often need to verify whether the AI agent’s responses are accurate and stable. To help with this, 3Chat.AI offers the Sandbox Testing feature, allowing you to test and debug the agent’s conversation flows in an isolated and safe environment.
1. Function Overview
The Sandbox Testing function provides a dedicated isolated environment for you to:
-
Simulate customer interaction scenarios
-
Validate the relevance and accuracy of the agent’s responses
-
Identify and fix potential issues
-
Optimize the knowledge base and conversation flows
This ensures that the agent configuration is correct before going live, improving the customer experience and conversation stability.
Important Notes
In the sandbox environment, only the knowledge base understanding ability is supported for testing.
AI tasks and handover to human agents will not take effect, and no new contacts will be created.
IM reply styles and continuous input handling configurations will not take effect in the sandbox.
If you wish to validate the Flow process in a production environment, please proceed to the Flow Space for testing.
2. Quick Operation Guide
2.1 Accessing the Function Page
1.Log in to the 3Chat Agent admin backend.
2.In the left navigation bar, click: 3Chat Agent → Testing → Agent Testing to enter the testing page.
2.2 Test Scenario Coverage
On the testing page, you can simulate customer input scenarios and check if the agent’s responses meet expectations.
Example:
- Input: How can I check my order information?
- Goal: Can the agent respond correctly and provide guidance on how to view order information?
2.3 Debugging and Optimization
- If the test reveals that the response is not ideal, you can go to the Knowledge Base or Fixed Scripts to adjust the content.
- After making adjustments, return to the testing page, click Reset Test, and re-enter the question for validation.
- After multiple rounds of debugging, ensure that the answer gradually improves and stabilizes.
3. Frequently Asked Questions (FAQ)
Q1: What should I do if the answer is incorrect during testing?
A: Please modify, update, or add the relevant Knowledge Base / Fixed Scripts and save the changes before retesting.
Q2: Can the input in the sandbox test be customized?
A: Yes, you can input any question to simulate various customer inquiry scenarios.
Q3: Will sandbox testing affect the production environment?
A: No. The sandbox testing is completely isolated from the production environment and will not modify any real data or task configurations.
4. Conclusion
The 3Chat.AI Sandbox Testing function provides a safe and flexible validation environment.
Before going live, you can ensure that the agent answers customer queries in a stable and accurate manner through multiple rounds of testing and optimization.
This not only reduces the risk during launch but also further enhances the quality and efficiency of customer service.