Key Features
- Safe Testing Environment: Any changes you make to your Agent’s identity, knowledge, or capabilities can be tested here without impacting connected channels (WhatsApp, web chat, etc.).
- Conversation Simulator: Interact with your Agent in real time to evaluate how it responds to sample questions or guided prompts.
- Predefined Prompts: Start quickly with suggested inputs such as:
- ¡Hola! ¿En qué me puedes ayudar?
- Cuéntame más sobre tu organización
- Restart Conversations: Reset the chat at any time to test a new flow from the beginning.
- Multilingual Support: Test your Agent’s responses across different languages (if enabled).
When to Use the Playground
Use the Playground whenever you want to:- Validate new knowledge entries before making them public.
- Confirm that capabilities (integrations, tools, functions) are responding as expected.
- Train your team by showing them how the Agent behaves in controlled scenarios.
- Debug issues spotted in production without needing to replicate them directly with live customers.
How to Access
- Go to Agents > Playground in the left navigation menu.
- Select the Agent you want to test (e.g., Customer Service – Ale).
- Type your query or click a suggested prompt to begin testing.
- Use the Restart Conversation button on the top right to clear the session and start fresh.
Best Practices
- Iterate Frequently: Test small changes in the Playground before pushing multiple updates to production.
- Compare Scenarios: Use similar inputs across sessions to measure improvements in responses.
- Document Learnings: Note issues or improvements needed and update the Agent’s knowledge or capabilities accordingly.
- Final Validation: Always confirm in Playground that your Agent’s responses align with your organization’s tone and service goals before enabling changes on public channels.
By using the Playground, your team can confidently experiment, refine, and optimize Agents—ensuring that what goes live has already been tested for quality and reliability.