Why AI matters at this scale

We Energies is a major regulated electric and natural gas utility serving customers in Wisconsin. Founded in 1896, it operates a vast network of power plants, transmission lines, and distribution infrastructure. As a utility in the 1,001-5,000 employee band, it has the capital resources and operational scale to invest in technology that can yield significant efficiencies, but must navigate a complex regulatory environment and legacy systems. AI is becoming critical for utilities to manage the energy transition, improve grid resilience, and meet rising customer expectations for reliability and service.

For a company of this size and vintage, the sheer volume of data from smart meters, supervisory control and data acquisition (SCADA) systems, and inspection reports is overwhelming for manual analysis. AI provides the tools to transform this data into predictive insights. The sector faces intense pressure to integrate intermittent renewable energy sources, reduce operational expenditures (OPEX), and modernize aging infrastructure. AI adoption is no longer a luxury but a necessity to maintain reliability, control costs, and support decarbonization goals within a regulated rate-of-return framework.

Concrete AI Opportunities and ROI

1. Predictive Grid Maintenance: By applying machine learning to sensor data from transformers, cables, and circuit breakers, We Energies can shift from time-based to condition-based maintenance. This prevents catastrophic failures that cause prolonged outages and require expensive emergency repairs. The ROI is direct: reduced capital expenditure on replacement equipment, lower OPEX for field crews, and improved reliability metrics that influence regulatory ratings and customer satisfaction.

2. Demand and Renewable Forecasting: Accurate short-term load forecasting is essential for efficient and cost-effective power procurement. AI models that incorporate weather, historical usage, and economic data can predict demand spikes, allowing for better scheduling of generation assets. Furthermore, forecasting wind and solar output optimizes the use of these resources, reducing reliance on fossil-fueled peaker plants and associated fuel costs, directly improving the margin on power supply.

3. Enhanced Customer Operations: Natural Language Processing (NLP) can automate the analysis of customer calls during storms, instantly categorizing issues and predicting outage locations. Computer vision applied to drone or helicopter imagery can rapidly assess storm damage. This accelerates restoration times, reduces the volume of calls to customer service centers, and improves public perception—a key intangible ROI in a regulated monopoly.

Deployment Risks for a Mid-Large Utility

Deploying AI at this scale involves distinct risks. First, integration with legacy operational technology (OT) is a major technical hurdle. Grid control systems are designed for safety and stability, not for rapid iteration with new AI software. Second, data silos between engineering, field operations, and customer service can cripple AI initiatives that require unified data lakes. Third, cybersecurity and regulatory compliance become more complex as AI systems interact with critical infrastructure; any breach or algorithm failure could have severe physical and reputational consequences. Finally, skill gaps may exist; attracting and retaining data scientists within a traditional utility culture requires clear career paths and executive sponsorship. Successful deployment requires phased pilots, strong collaboration between IT and OT teams, and a focus on use cases with unambiguous regulatory and business alignment.