AI Energy & Utilities Agent

Description

We operate about 60 commercial buildings and our energy data arrives from four different metering systems, two of which still deliver CSV by email. Consumption anomalies get noticed on the monthly bill, which is roughly three weeks too late to do anything about. We want an agent that watches the daily feeds, flags anomalies against the building's own baseline rather than a fleet average, and writes an explanation an facilities manager can act on — "site 14 overnight base load is up 30% since the 8th, HVAC schedule change on the 7th" beats a red dot on a chart. Weather normalisation is required. Without it half the flags will be January.

• Ingest and normalise all four metering feeds, including the emailed CSVs, with per-source health checks
• Build per-building baselines with weather normalisation and occupancy awareness
• Generate written anomaly explanations linked to the underlying interval data
• Alert the responsible facilities manager per site, not a shared inbox
• Report monthly on flagged anomalies, actions taken and estimated consumption avoided

• Ingestion pipeline across all four metering systems with feed monitoring
• Weather-normalised anomaly detection per building
• Written alerting routed by site ownership
• Monthly performance report