Session Abstract

Building a Copilot Studio agent has become almost too easy — a working demo in under an hour. But what happens after that first success? Who owns the agent in production? What prevents it from exposing sensitive data or triggering unintended actions? Agents introduce a new class of governance challenges: they act autonomously, combine connectors dynamically, and don’t always behave deterministically. That makes traditional Power Platform ALM and DLP approaches necessary — but not sufficient. In this session, I’ll walk through the lifecycle of an agent based on a real client engagement — from first build to production operation — and where things typically go wrong. We’ll cover: Environment strategy and Managed Environments for Copilot Studio ALM pipelines using managed solutions across dev, test, and prod DLP policies for agents — and how they differ from traditional apps and flows Security roles, ownership, and the “what happens when the maker leaves” problem Licensing strategies (pay-as-you-go vs. credit packs) and their governance impact You’ll leave with a practical blueprint for managing agents responsibly — before your organization creates its own shadow AI problem.

Gerjon Kunst

'm Gerjon, an incurable tech optimist with 25+ years of experience pushing the boundary between "can't" and "can" and then automating the boring bits. Today, I help organizations move from AI hype to real business impact. As AI Lead at Interstellar Group, my mission is simple: close the gap between AI ambition and AI results. I bring together AI expertise across the Interstellar ecosystem and help turn AI opportunities into practical solutions that deliver measurable business value. Alongside this role, I serve as CTO at RawWorks, where I help shape practical AI strategies, architectures, and ways of working that customers can actually use. As a Microsoft MVP in both M365 Copilot and Copilot Studio, I actively help organizations understand, adopt, and scale AI solutions that create real outcomes. A Microsoft MVP in AI, speaker, podcast host, and lifelong learner, I believe the gap between AI hype and AI results is closed through simplicity, trust, and execution. 🎤 Speaker Regular speaker at conferences, community events, and user groups, translating complex AI and technology trends into practical insights and actionable guidance. 🎙️ Podcast Host Co-host of Impact of AI: Explored, where together with James O'Regan I explore the opportunities, challenges, and real-world impact of AI.