Skip to main content
The Simulator allows you to test AI chat or email interactions. It has two tabs: the Playground for ad-hoc manual conversations, and Tests for automated test scenarios with assertions and AI evaluation. You can also use the simulator to test changes to your AI Agent’s behavior, knowledge, and actions before publishing them.

How to Start a New Conversation

Click + New Conversation. You will see a pop-up window with two channel options:
  • Chat
  • Email

Chat Channel

1

Select Chat

Select Chat as your channel.
2

Enter email (optional)

Optionally, enter an email address if you want to link the chat to a specific user.
3

Include unpublished changes

Toggle the Include unpublished changes option to either test your changes before publishing them or test the live configuration.
4

Start

Click Start. You’ll be redirected to the Playground, where you can test real-time interactions between the AI and a user.
Chat Simulator Setup

Email Channel

1

Select Email

Select Email as your channel.
2

Enter recipient

Enter the recipient’s email address.
3

Add subject

Provide a Subject line.
4

Add CC (optional)

Optionally, add CC Address.
5

Include unpublished changes

Toggle the Include unpublished changes option to either test your changes before publishing them or test the live configuration.
6

Start

Click Start. You’ll be redirected to the Playground, where you can send and receive simulated email exchanges.
This feature is ideal for validating Specialist behavior and fine-tuning instructions before going live.
Email Simulator Setup

Playground Interface

Once inside the Playground, you can:
  • Send test messages as a user or agent.
  • Observe how the Specialist responds to inputs.
  • Test handoff conditions, response logic, and conversation flow.
Tickets created via the simulator will not be shown inside your inbox.

See also

  • Automated Tests: Create automated test scenarios with assertions and evaluation
  • Specialists: Create topic-specific AI behavior
  • Knowledge: Add knowledge sources for the AI
  • Actions: Configure automated and manual actions
  • Publishing: Publish changes to your AI Agent’s behavior, knowledge, and actions