2026 Community of Human and Organizational Learning 32nd Annual Conference

Agentic AI is gaining attention as a new class of systems that move beyond analysis to actively observe, interpret, and support operational decisions. This session begins with a brief, practical overview of what agentic AI is, and what it is not, before turning quickly to where it can be applied today.
The focus is on real-world use cases across three areas that are relevant to many of the practitioners in the attendance: safety, HOP, and RCA. Taken together, these three areas reinforce each other: connecting everyday operations, learning from events, and sustained risk reduction. Agentic AI can help tighten the connection by stitching signals, context, and actions into a continuous cycle of improvement.
In safety, agentic systems can sit on top of existing safety data and management systems, continuously monitoring signals, applying predictive models, and delivering plain-language guidance to frontline teams. Rather than static dashboards, these systems combine data, models, and agent-based interaction to help identify emerging risk, support coaching, and shift organizations from reactive reporting to forward-looking prevention.
In Human and Organizational Performance (HOP), agentic AI can make operational load more visible, surfacing signals of demand, constraint, and tradeoffs—and help identify conditions where the system is nearing its capacity to fail, not just its capacity to perform. By highlighting these conditions in context, agents can support better metrics and the subsequent conversations around system design, work-as-done, and organizational learning.
In root cause analysis (RCA), agentic approaches can help address one of the biggest constraints: fragmented and incomplete information. Agents can automatically assemble timelines, pull together reports, logs, and records, and surface contributing factors more quickly, reducing the manual effort and bias that often lead to shallow conclusions. This allows teams to move beyond symptom-level fixes and focus on underlying system conditions that drive repeat events.
The session also takes a grounded view of current limitations. Many agentic AI solutions are oversold, particularly in their level of autonomy and reliability. The most effective applications today are those that augment human judgment, operate within clear boundaries, and are tied closely to real operational context.
Attendees will leave with a clear understanding of where agentic AI is delivering value now, and how to evaluate opportunities without getting caught up in the hype.