Avangrid is expanding artificial intelligence across its U.S. utility operations, moving beyond isolated pilots toward everyday applications in grid reliability, asset management, field operations and customer service. The company announced the latest expansion on September 14, 2026, covering AI, advanced analytics, digital grid capabilities and responsible data governance. Avangrid operates seven electric and natural gas utilities serving more than 3.4 million customers in New York and New England, with operations across 25 states and more than 11 GW of installed generation capacity.
One of the most immediate applications is AI-supported inspection of electricity infrastructure. Avangrid crews are using drones to capture aerial information about grid assets, with analytics applied to the resulting data. The approach is intended to identify potential problems earlier and help maintenance teams focus on assets requiring attention instead of relying only on conventional inspection schedules.
The company has also been developing proprietary machine learning capabilities for grid planning and asset health. Its Predictive Health Analytics system assesses substation equipment such as circuit breakers using factors including equipment age, usage, manufacturer information and maintenance records to support replacement decisions before failures occur. GeoMesh divides service territories into smaller geographic areas and analyzes data including weather conditions, outage history, population and vegetation to assess grid performance under normal and severe-weather conditions. HealthAI uses high-resolution images of distribution infrastructure to identify equipment such as poles, transformers, crossarms and wires, with the longer-term objective of detecting physical deterioration and potential hazards proactively.
Avangrid is also applying generative AI directly to field work. Its First Time Right Autopilot, developed with Amazon Bedrock, allows technicians to interact with an AI assistant through voice or text on a mobile device when troubleshooting wind turbine problems. The system uses Avangrid's operations and maintenance engineering knowledge to provide diagnostic guidance, supporting documents and instructional videos. The technology was initially deployed at wind facilities in Iowa and New York and was designed to help technicians resolve problems more quickly and return turbines to service.
Another emerging area is agentic AI. Avangrid's 2026 Agentic AI program identifies Ask E&P as an example of an AI assistant supporting environmental and permitting work. The company has also outlined the concept of Ask Avangrid, which could bring together department-specific assistants across areas such as power and operations, information technology, and environmental and permitting functions. This points toward AI being used not only to retrieve information, but also to coordinate knowledge and workflows across different parts of the utility.
Customer-facing applications form another part of the strategy. Avangrid is using voice-of-the-customer analytics to examine interactions, identify sentiment and recognize recurring service issues. Its networks have also developed AI-supported customer tools, including Ava, which was enhanced to provide estimated power-restoration times during outages. This connects AI deployment directly with outage communication and customer experience rather than limiting it to back-office applications.
Avangrid's approach is accompanied by a formal AI governance framework. Its April 2026 AI policy establishes principles covering responsible use, transparency, security, legal compliance, human intervention and protection against discriminatory outcomes. The company also joined the Electric Power Research Institute's Open Power AI Energy Consortium in 2025, participating in work on domain-specific AI models, datasets, testing environments and responsible deployment for energy systems.
The combination of predictive asset analytics, drone-based inspection, generative AI for technicians, customer intelligence and emerging agentic assistants shows how Avangrid is building AI into multiple layers of utility operations. Rather than depending on a single technology, the company is applying different AI capabilities to equipment condition, weather-related risk, maintenance decisions, field troubleshooting, permitting and customer communication as it works toward a more data-driven and adaptive grid.