
This process walks you through testing your updated AI agent using the built-in chat simulator in the Retell AI dashboard, verifying its responses, making any needed adjustments, and publishing it when it's ready to go live.
Navigate to your Retell dashboard and open the agent editor for the specific AI agent you want to update. On the right-hand side of the screen, locate the testing simulator panel, which serves as a dedicated environment to converse with your agent before publishing it.
In the test simulator, select how you would like to interact with the agent. You can test the agent using audio or chat.
For this walkthrough, select the chat testing functionality.
In the chat simulator, type messages as a customer would and test different scenarios.
For example, you might ask, “Hey, what time are you open?” or other frequent customer questions. Observe how the agent responds to ensure your recent changes are behaving as expected and that it's pulling the correct information from the knowledge base, triggering functions correctly and responding in a natural way that you're happy with.
If you notice anything incorrect, missing or off-brand in the agent’s responses, make further adjustments in the prompt configuration, the connected functions, or the knowledge base, wherever the issue originates.
Repeat testing in the simulator until you are satisfied that the agent is behaving correctly and reliably.
Once you are happy with the test results and ready for the agent to go live, click the Publish button at the top of the agent editor.
This action will publish the agent to the contact number so that real customers can begin interacting with it. After publishing, your agent is live and ready to handle incoming conversations.