The useful AI work in buildings starts with operator context: noisy telemetry, alarms, comfort complaints, energy targets, and maintenance constraints.
The opportunity is not only smarter analytics. It is turning raw operational complexity into decisions a building team can actually trust.
Most operators are not waiting for another dashboard. They already have dashboards, alarms, trend logs, BMS screens, work orders, tenant complaints, and energy reports. The problem is that these signals arrive fragmented. AI becomes useful when it can connect them into an explanation: what is happening, what changed, what is likely to happen next, and what action is worth taking.
Energy optimization is the obvious entry point because it has measurable value. The product challenge is to avoid treating optimization like a black box. Operators need to understand why a recommendation exists, what comfort tradeoff it implies, and whether the system is learning from real building behavior or simply applying a generic rule.
Fault Detection and Diagnostics is similar. A useful FDD experience does not just say “fault detected.” It ranks urgency, shows confidence, connects the fault to affected equipment, and helps a team decide whether this is a commissioning issue, a controls issue, a sensor issue, or a maintenance issue. That workflow layer matters as much as the model.
GenAI can help with the softer work around operations: summarizing alarms, drafting maintenance notes, explaining sequences, creating commissioning checklists, and helping new operators understand unfamiliar systems. The best use cases reduce cognitive load without pretending the system knows more than it does.
For me, the product bar is simple: AI should make the operator feel more in control, not less. If it cannot explain itself, fit into the daily workflow, and earn trust through small useful moments, it is not ready to be more than a demo.