How to Build an AI Agent Workflow with Camunda

Learn to build an AI agent workflow that integrates human oversight and REST API calls using Camunda 8.8.0-alpha8 and BPMN (Business Process Model and Notation) in this 13-minute tutorial. Discover how to create a BPMN diagram in Camunda Modeler and implement an AI agent using BPMN ad-hoc sub-processes, which provide the flexibility needed for dynamic elements within deterministic processes. Master the configuration of AI agents and learn to add REST APIs or other available tools to expand your agent's capabilities. Explore human-in-the-loop integration through BPMN user tasks, ensuring proper oversight to prevent AI unpredictability, errors, and hallucinations that can lead to data misuse and compliance challenges. Understand how to instruct AI agents to create outputs using "fromAi()" functions and create linked user forms within Camunda Modeler. Follow the complete deployment process to run your AI agent workflow in Camunda 8.8.0-alpha8 and implement user feedback integration for continuous improvement. The tutorial emphasizes industry best practices for AI agent implementation, including clear expectations, detailed toolboxes, and straightforward escalation paths to ensure trustworthy and reliable AI automation within enterprise environments.

Syllabus

00:00 – Welcome + Introduction
00:24 – How to build a BPMN diagram in Camunda Modeler
01:16 – How to add an AI agent in the form of a BPMN ad-hoc sub-process to your BPMN model
01:49 – How to configure an AI agent
03:47 – How to add a REST API, or any available tool, for your AI agent to use
05:46 – How to integrate human-in-the-loop, using a BPMN user task
07:21 – How to tell an AI agent to create an output using “fromAi”
08:49 – How to create and link a user form in Camunda Modeler
10:11 – How to deploy and run the AI agent workflow in Camunda 8.8.0-alpha8
11:30 – How to integrate user feedback for your AI agent.