Publishing and Deploying Worker Agents
Once your Worker Agent is tested and ready in Studio, you will need to publish and deploy it to make it available for production use.
The steps for publishing and deploying a Worker Agent are exactly the same as for Conversational Agents and other agent types in Sema4.ai.
Deployment Overview
The deployment process has two stages:
-
Publish from Studio:
Once your Worker Agent is finalized and tested in Studio, you publish it to the Control Room. This will move your Worker Agent from development to your organization's Control Room, making it discoverable for production. See how to publish your agent for detailed steps with screenshots -
Deploy in Control Room:
In the control room, your admin or DevOps team can deploy the published Worker Agent to a runtime environment. This involves selecting the appropriate compute resources, configuring scaling options, and setting up monitoring. See how to deploy your agent in the Control Room.
What happens after deployment?
Once deployed, your Worker Agent is live. You can begin submitting Work Items via the Work Item API through integrations, or schedulers, or triggers. Learn how to add Work Items here. Monitor execution, review outcomes, and trigger human-in-the-loop actions for cases requiring review.