AI in Electric Utilities

ESB Networks, Ireland's electricity distribution operator with about 2.4 million customers, has started a five-year programme with eSmart Systems to inspect up to 10,000 grid structures. Drones capture images under standard protocols, computer vision flags defects, an analyst reviews every finding, and each structure gets a record of image, location, condition and recommended action.

AEP Ohio flew drones in 2025 where inspectors would typically work on the ground, finding 150-plus tier-one issues, mostly thermal anomalies seen by the drone's thermal sensor. One person then spent 500-plus hours reviewing 400,000 to 500,000 images, so for 2026 the utility is looking to a Levatas integration.

Ausgrid, the largest energy network on Australia's east coast, piloted a KPMG guided-buying accelerator built on KymChat and Microsoft Azure OpenAI, with selected Ausgrid procurement policies and rules embedded so staff get policy-aligned guidance. Ausgrid reports fewer errors, better efficiency and early cuts in manual effort; no figures were published.

ComEd's grid analytics team built a computer vision model that analyzes drone images of wooden pole tops and assigns each pole a health score to rank replacement priority. The model has reached 86 percent accuracy against human expert agreement, and ComEd calls the approach emerging, with potential to speed up maintenance and reduce outages.

Dutch energy supplier Essent runs AI-first customer service with NiCE Cognigy: an Agent Copilot in Microsoft Teams trained on 1,100+ articles now handles 85% of former helpdesk queries, saving EUR 2 million a year; AI handles all incoming chats first and fully automates 55%; and a voice agent collects meter readings, reaching 22,000 customers in two weeks.

Bayernwerk Netz, a German grid operator owned by E.ON, worked with E.ON Digital Technology to build Bayernwerk AI Agents on the Databricks platform, a field operations assistant that compiles specifications and safety procedures into source-linked briefs. The tool is in a pilot with 60 users, and a rollout to about 4,000 business users is planned.

National Grid has moved sees.ai's autonomous drone inspections from R&D into business as usual on its electricity transmission network in England and Wales. Drones piloted from a central control room fly beyond visual line of sight close to live pylons and conductors, after a four-year innovation programme, with AI and machine learning used to optimise the inspection data.

National Grid ESO, the UK electricity transmission system operator, partnered with Open Climate Fix and now uses an AI solar forecasting model in its control room — combining real-time satellite imagery with weather and solar generation data to deliver three times the accuracy of the prior approach, with control room engineers using the forecasts to reduce balancing costs and emissions.