Deploying Agentic AI Workflows in Clinical Environments Learning Lab

As AI moves beyond single-task tools, agentic AI—systems that plan, act, and coordinate across tools—is quickly becoming essential in medical research, education, and clinical operations. Yet many radiology and imaging professionals lack accessible, hands-on ways to experiment with agents in real workflows. This learning lab addresses that gap by showing attendees how to deploy and use agentic workflows responsibly, without requiring coding or advanced infrastructure. 
 
Attendees learn how agentic AI can automate and orchestrate common radiology and informatics tasks, including PACS integration, scheduling, dataset preparation, alerts, and clinical knowledge access. Participants work with n8n, an open-source automation framework, to build practical workflows they can adapt to local or cloud environments. By the end of the lab, attendees leave with reusable workflow templates, a working understanding of agent-based design patterns, and concrete skills they can immediately apply to research, education, and clinical operations. 
 
The lab combines short conceptual framing with guided, hands-on exercises. Attendees follow step-by-step setup instructions, then actively build and run workflows that mirror real clinical scenarios—connecting to PACS via DICOMweb, integrating scheduling systems, anonymizing datasets, triggering notifications, and creating a retrieval-augmented (RAG) chatbot for clinical Q&A. Time is reserved for discussion, adaptation to institutional constraints, and sharing extensions so participants can continue developing agentic workflows after the session.  

Objectives 

• Describe the role of agentic AI in automating and coordinating workflows across medical research, education, and clinical operations.
• Implement no-code agentic workflows using an open-source automation framework to support imaging, scheduling, and data management tasks.
• Build AI-powered assistants that retrieve and answer questions from clinical notes, reports, or operational documents.
• Deploy an open-source automation framework locally or in the cloud to support agent-based workflows.
• Create automated workflows for PACS integration, dataset anonymization, scheduling, and critical alert notifications.
• Use a retrieval-augmented generation (RAG) pattern to develop a chatbot that supports clinical, educational, or operational Q&A.