The self-learning building; agentic AI in building management

This session focuses on the operational side of AI, and we look ahead to a future where building management is placed in the hands of AI agents. Using concrete practical examples, we outline how AI is already actively predicting and making decisions in building management today. Consider predictive maintenance, where computer vision detects malfunctions before they occur; AI agents that adjust HVAC systems in real time based on occupancy, weather, and energy prices; and LLM-driven interfaces that allow facility managers to communicate with their building management system in plain language.

In addition to the technology, the less glamorous side also receives attention: governance, data quality, and adoption. After all, most AI projects, including those in building management, do not fail due to technology, but because of everything surrounding it. Drawing on his work at Datacation—where organizations are helped to bring AI from proof-of-concept to production—Douwe Groenevelt shares a down-to-earth perspective on what works, what does not (yet) work, and where the real ROI lies.