Playground

The Playground allows you to test your AI Agent before making it available in production.

It provides a safe environment to validate how the agent responds, whether it uses the correct knowledge and skills, and whether its behavior matches your intended configuration.

This is the recommended place to test your agent after configuring:

• Agent Details

• Knowledge

• Skills

• Channels

---

What You Can Do in Playground

In the Playground, you can interact with your AI Agent directly and review how it behaves in real conversations.

Use it to:

• Test responses to user questions

• Verify whether the agent uses the correct knowledge

• Check if the correct skills are triggered

• Review the agent’s reasoning for each response

• Identify changes needed before production use

---

How It Works

1. Enter a message in the chat input

2. Send the message to the AI Agent

3. Review the response

4. Expand or inspect the reasoning shown for that answer

5. Adjust the agent configuration if needed

This helps you understand not only what the agent answered, but also why it answered that way.

---

Reasoning Visibility

For each answer, the Playground displays the agent’s reasoning.

This helps you review:

• How the agent interpreted the user input

• Whether it used knowledge, skills, or only instructions

• Why a specific response was generated

• Whether the behavior aligns with your expectations

This is especially useful when troubleshooting:

• Incorrect answers

• Missing knowledge usage

• Unexpected skill behavior

• Instruction issues

---

When to Use Playground

Use the Playground whenever you:

• Create a new AI Agent

• Update instructions

• Add or change knowledge

• Add or change skills

• Want to validate behavior before enabling a channel

• Need to troubleshoot responses

---

Best Practices

• Test both simple and realistic user prompts

• Try expected use cases and edge cases

• Verify that the agent stays within its intended scope

• Review reasoning to confirm correct decision-making

• Make changes and retest before going live

---

What’s Next

Once your testing is complete and the agent behaves as expected, you can continue by enabling production channels such as:

• WhatsApp

• Microsoft Teams

• Webhook

• Mate Chat

---

Summary

The Playground is your testing and validation environment.

It helps you:

• Check agent quality

• Review reasoning

• Catch issues early

• Improve the agent before production use