Human-in-the-Loop AI in Power Platform: Designing AI Around Real Human Behavior

AI is adopted inside existing habits and decision patterns, not inside idealized process models. This session focuses on how to design AI experiences that work with how people already behave, rather than against it. We’ll start conceptually by examining why Human-in-the-Loop (HITL) is less about control and more about fitting AI into existing human judgment patterns, SME instincts, and trust boundaries. From there, I’ll demonstrate a simple example solution built on the Microsoft Power Platform that shows how human review, feedback, and knowledge capture can be embedded directly into an AI-assisted process using Model-Driven Apps, Dataverse, Agents and Agent flows. The example showcases on how human corrections and decisions into reusable organizational knowledge. The goal is not to automate humans out of the loop, but to design AI systems that respect how people actually work — where they hesitate, where they override, and where they rely on context. This session equips you with practical patterns for building AI-powered business processes that learn from human behavior instead of trying to replace it.