Server Agents
Server agents run centrally in IB-X. Design them in Agent Builder and execute and monitor them through the AI Command Center.
Prerequisites
- AI Command Center access with permission to create, edit, test, publish, and run agents.
- The licensed Agent Builder feature enabled by an administrator.
- Connections and service access required by the selected tutorial. Endpoints must be reachable from the server execution environment.
- For AI and conversation examples, a configured AI model connection. Use your provider connection or Integration Gateway managed AI services, which consume IB-X Currency.
Shared design and execution steps
- Click New Agent, enter a name and description, and click Create.
- Add activities in Agent Builder. Connect an activity's output port to the next activity's input port.
- Configure activities and bind values using the Process Data drawer. Insert the references provided by the designer rather than assuming a variable path.
- Save and test using a Debug Automation. Inspect activity results and logs.
- Publish the tested agent. Use Run for manual execution, or the configured trigger or channel for event-driven execution.
Follow Agent Life Cycle for detailed testing and publishing instructions. Time triggers are registered when published, not when a draft is saved.
Choose your learning path
| Path | Starter example |
|---|---|
| Automation Agents | Retrieve an order through an API |
| Workflow Agents | Collect a request and wait for human approval |
| Conversation Agents | Build a web chat assistant with a tool |
| AI Agents | Classify customer feedback |
These are usage patterns within the same agent framework. For example, a workflow can use AI to classify a request, ask a person to review it, and call an API after approval.