STORM-SHIFT — Storm restoration orchestration agent

STORM-SHIFT — Storm restoration orchestration agent

Problem: Storm-related outages cost the U.S. billions annually. Legacy SCADA and outage-management systems remain siloed, crews get dispatched on stale damage pictures, mutual-aid coordination is manual (risking roster overlaps), and customer communication is batch SMS blasts sent hours late — while regulatory penalties for poor reliability keep rising. Restoration speed is decided in the first hours, when situational awareness is worst.

Idea: Add a watsonx Orchestrate agent that fuses pre-storm asset baselines (vehicle-mounted inspection data such as Noteworthy AI's grid-intelligence program), real-time storm reports, and outage telemetry into a living damage map; builds restoration queues prioritized by customer criticality; verifies crew certifications on the fly and assembles mutual-aid rosters without overlaps; and produces timestamped, FEMA-ready documentation so eligible disaster-assistance costs do not go unreimbursed. Human dispatchers approve every crew assignment — the agent proposes the plan, people send the trucks.

Value: Restoration moves at the speed of the damage map instead of the rumor mill (the EY–Eversource pattern avoided roughly 40,000 customer outages in two months; KYRO AI users report SAIDI down 20% and SAIFI down 15%), mutual-aid chaos becomes a coordinated roster, and timestamped evidence protects the utility's disaster-assistance claims.