LLMs are only one category within a much broader AI landscape. In many energy and industrial use cases, other forms of AI have been delivering value for years.
Machine Learning (ML)
Machine learning models learn patterns from historical data to make predictions or classifications. These models are commonly used for:
• Load, price, and demand forecasting
• Anomaly detection in trading, operations, or asset performance
• Risk scoring and scenario analysis
Unlike LLMs, ML models are typically narrow, highly specialized, and deeply embedded into operational systems.
Computer Vision
Computer vision applies AI models to images, video, and sensor data. In energy and utilities, this includes:
• Asset inspection using drones or cameras
• Leak detection and infrastructure monitoring
• Safety compliance and environmental monitoring
Optimization and Prescriptive AI
These models do not just predict outcomes but recommend actions. They are used in areas such as:
• Grid optimization
• Scheduling and dispatch
• Portfolio optimization and constraint-based planning
Agentic or Autonomous AI
This emerging category combines LLMs, ML models, and business rules to create systems that can plan tasks, take actions, monitor outcomes, and adjust behavior with limited human input. According to McKinsey & Company, agent-based AI represents a significant shift from tools that assist humans to systems that can execute end-to-end workflows.
This is widely viewed as the next major evolution of enterprise AI, but it also introduces new governance, risk, and control considerations.
