The Workflow Behind Agentic Course Development

When Sean Rugge set out to turn a lengthy Marine Corps mountain warfare manual into an interactive online course, he wanted to find out what AI could genuinely contribute to course development. The result was a proof of concept that combined AI-generated course structure, learning activities and media with an agentic browser that handled some of the repetitive work inside the learning management system.

In this episode of Tech Transforms, Sean, a senior educational technologist at Marine Corps University, walks Carolyn Ford through the course and the decisions behind it. He explains how AI helped organize source material and build a branching scenario—and how his own review caught an omitted chapter and unreliable claims about time savings. A Marine subject matter expert provided another essential check on the training content.

Carolyn and Sean also discuss a harder question: if an AI agent can take a course for a student, what should the course ask students to do? Sean makes the case for assignments that require learners to direct, assess and improve AI-generated work. It is a practical look at what AI can speed up, what educators still need to verify, and how training can develop critical thinking alongside AI skills.

Show Notes:

LinkedIn: https://www.linkedin.com/in/sean-rugge-1ba47a171/

The Phantom Tollbooth by Norton Juster.

Sean’s github repo with 2 of the prompts I use for analysis: The session analyzer and the discussion analyzer. https://github.com/seanrugg/local_lid/tree/main/prompts