How energy teams turn theft detection into governed action with Genie and AI business processes

Energy theft in Great Britain costs consumers over £1.4 billion annually, with only 40% of cases detected despite existing ML models flagging suspicious accounts. The challenge has shifted from detection to operationalizing insights into a unified workflow that accelerates investigations, recovery, and reporting. Databricks’ platform integrates detection, triage, field dispatch, and recovery into a single governed loop, reducing delays that prolong financial losses and safety hazards.

The workflow begins with an AI-generated case summary explaining why an account was flagged, paired with a dispatch-ready report including evidence checklists and safety notes. This replaces manual reviews with consistent, plain-language assessments, enabling analysts to prioritize investigations and prepare field teams efficiently. The Databricks App maintains live case state in Lakebase, allowing real-time updates and immediate tracking of recovered revenue.

Executive visibility is enhanced through Genie One, which answers leadership questions in plain English using governed metrics like revenue recovered and precision rates. An Agent Bricks Multi-Agent Supervisor automates month-end reporting, ensuring traceable, board-ready outputs. Governance remains central, with Unity Catalog tracking data lineage, applying access controls, and ensuring compliance with Ofgem’s Data Access and Privacy Framework.

The approach extends beyond energy theft, offering a repeatable pattern to operationalize model outputs for fraud, maintenance, or churn interventions. By embedding AI-driven insights into governed business processes, organizations can act faster while maintaining regulatory compliance and data security across the workflow.