Opening an Agent from the Agents list displays its Agent workspace. This is the primary operating area for training, testing, reviewing, publishing, and improving that specific Agent.
The workspace header can display:
Use the status information before assuming the Agent's latest configuration is available to website visitors.
The Agent workspace calculates setup readiness using three operational steps.
This step becomes complete after the Agent has at least one successfully indexed Knowledge Source.
If it is incomplete, the workspace directs you to add a URL or file so the Agent can answer from real content.
This step becomes complete after the Agent has been tested in the Playground and a Playground Conversation contains messages.
The purpose is to verify realistic questions, tone, accuracy, and fallback behavior before launch.
This step becomes complete when the Agent has been published.
Publishing makes the Agent ready for website installation.
The workspace displays progress as a count out of three, such as 2/3, together with a readiness message and progress bar.
When all three steps are complete, the Agent can display Ready for installation.
The Agent workspace automatically chooses a next action based on readiness:
The Agent workspace provides direct access to core operating areas including:
Additional Agent tools can include:
The Agent workspace also displays the current System prompt for reference. If none has been defined, the workspace indicates that one can be set in Settings.
Use the Agent workspace as the operational home for one Agent. Follow setup progress in order, use the provided next action, and return to the workspace whenever you need to move between training, testing, publishing, and live-performance review.