March 19th, 2026 @5:30 PM – Randy Soper on AgentOps Concepts and Shadow Agents

Register: This is a Hybrid meeting. A dinner meal will be served (Meal menu TBD). REGISTER HERE for IN PERSON & ONLINE Zoom, Location Marymount University Ballston Center 1000 N Glebe Rd, Arlington, VA 22201. This will be in room: BALL 4004. Metro Station Accessible. Registration closes Wednesday, 03/18/26.

Abstract: We are witnessing a fundamental shift in Generative AI: the transition from “Chatbots” that talk to “Agents” that act. While national policy debates focus on the existential risks of Artificial General Intelligence (AGI), a more immediate and tangible threat is emerging within the enterprise: Shadow Agentic AI. In this session, we will move beyond the hype to explore the emerging reality of autonomous agents. We will examine the emerging discipline of AgentOps, the lifecycle management of agentic behavior, and how it differs from traditional MLOps. We will look at agentic AI in its recent historical context and develop some basic mental models that help us to understand where AI is and where it is likely going. We will consider both the state of the art – capabilities and gaps – as well as some likely directions for continued agentic AI development and adoption. This talk will help to identify some key emerging organizational risks around agentic AI and what security professionals need to do to prepare and to position their organizations.

About the Speaker: Randy Soper is a nationally recognized expert in enterprise-scale artificial intelligence, with over 25 years of experience in modeling and simulation, digital transformation, and data science spanning government, industry, and academia. He currently serves as Senior Director and Federal Data & AI Lead at Slalom, where he helps federal agencies harness data and AI to drive mission impact and modernization.
Previously, Randy was the Chief of Artificial Intelligence at the U.S. Internal Revenue Service, where he led the agency’s AI technology modernization, architecture, and governance. He has also contributed to the U.S. Intelligence Community’s responsible AI standards and advised on AI risk and security supporting NIST and MITRE frameworks. A thought leader and frequent speaker on AI infrastructure, strategy, assurance, and security, Randy serves on the IEEE AI Standards Committee, the advisory board of AI 2030, and the editorial board of CDO Magazine.
Randy holds a master’s degree in mechanical engineering from Virginia Tech and serves on the department’s advisory board.