Use AI Agents in a Process

Integrate your AI agents directly into your business workflows by calling them from a process model. This allows you to automate tasks that require generative AI, passing contextual data from your process and receiving structured outputs. You can run an AI agent from any process model using the Execute AI Agent smart service.

For foundational concepts, see About AI Agents. For configuration guidance, see Create and Configure an AI Agent.

To use an AI agent in a process model:

1. In your process model, drag the Execute AI Agent smart service from the node palette onto the canvas.
2. Double-click the smart service to open its properties and go to the Data tab.
3. On the Inputs tab, provide the AI agent's UUID and necessary inputs.
4. On the Outputs tab, configure process variables to store the agent's outputs.
5. Click OK.

See the Execute AI Agent Smart Service page for detailed information on all node properties, outputs, and usage considerations.

For testing your process integration, see Monitor AI Agents. For troubleshooting process-related issues, see Troubleshooting AI Agents.

Process agents only

The instructions on this page apply to process agents only. Process agents are designed to run within process models using the Execute AI Agent smart service.

To deploy a chat agent in an Appian interface instead, use the agent chat component (a!agentChatField). Chat agents cannot be executed via the Execute AI Agent smart service, and process agents cannot be used in the agent chat component.

Feedback