Test & Review
Test & Review is where you exercise the agent before and during launch. You can simulate conversations, chat through the real web widget, and review draft replies in a queue.
Simulator
Section titled “Simulator”Run sample conversations against the agent to see how it responds. You can build and save scenarios, replay them, and compare runs — useful for confirming a change behaves the way you intended.
Widget test
Section titled “Widget test”The widget test talks to the agent over the real web chat channel, just as a customer would. Use it to sanity-check the live experience end to end, including attachments and captions.
Review queue
Section titled “Review queue”While the agent is in shadow mode or a partial launch, its replies can be held as drafts for you to approve, edit, or reject before they’re sent.
Related
Section titled “Related”- AI agent overview — how the agent decides what to do.
- Launch & history — phases, versions, and results.
- The Inbox — where approved replies appear.