The electric energy industry, like many, is on a learning curve when it comes to applying artificial intelligence (AI) to their businesses. That said, extensive research, pilot programs and commercially operational use cases for AI applications are on the rise in the power and utility sector. As the adoption of AI within the sector increases, it is anticipated that power and utility companies will leverage it and see improvements in overall asset, employee and financial performance.

In recent years, the utility industry has seen an uptick in AI adoption and use cases. However, many are still evaluating AI and trying to fully comprehend the technology’s capabilities and decide which areas within their organizations are best suited to deploy AI technology and models. Whether applied in operations, at the enterprise level or in the back-office, power and utility companies using AI applications are beginning to realize positive results. With a foundational AI strategy in place, organizations in the sector are expected to produce positive business, operational and financial outcomes.

The following are key areas within the power and utility sector that are implementing AI with expectations for a return on investment.

Asset performance

A solid use case for AI in the power and utility sector is in the critical area of asset performance. Assets and their operational performance are the backbones of the electric power industry sector. AI models are being deployed to increase the operational and financial performance of their asset portfolios. Whether in generation, transmission or distribution, the assets in core areas of the power and utility sector’s value chain can improve asset availability, optimization and life cycles when leveraging AI.

For example, in its 2023 Worldwide Utility MarketScape, one global provider of market intelligence and advisory services predicts that by the end of 2023, 60% of competitive power generators and traders will have AI-powered forecasting capabilities in production to help improve day ahead demand and wholesale price forecasting accuracy by more than 15%. Increasing the accuracy of electric demand and price forecasting provides power and utility companies with the potential to experience sizeable financial gains through the use of AI models. Utilities and competitive power producers have been using neural networks for years to better predict demand forecasts. Building, running and continuously training even more sophisticated AI models to increase the accuracy of electric demand forecasting models by analyzing years of historical data – supported by several variables such as time of year, time of day, similar heating and cooling degree days and more – can vastly improve their ability to meet demand obligations in the most economical manner.

Integrated resource planning

Many regulated utilities are required to submit integrated resource planning data to public utility commissions (PUCs) to prepare for the expected demand growth and generation sources that will be needed in future years to reliability serve their customers. AI can be used to create forward-looking model simulations to better predict load growth and energy supply which provides utilities with better forecasts of not only power supply and demand but can accurately forecast capital and operational expenditures as well. Historical and forecasted data on supply and demand fundamentals is essential when producing accurate capital and operational spending forecasts.

For example, many utilities are beginning to build out digital twin models, which are digital replicas of their utility networks and assets. These digital twins that are being deployed are leveraging AI capabilities to better understand both traditional utilities connected in front of the meter and emerging behind-the-meter forecasts of energy resources available in their footprints for years to come. With the expected accelerated growth of distributed energy resources (DERs) such as electric vehicles, rooftop solar and energy storage, it is becoming more difficult for utilities to have a clear and confident view of what capital and operational investments are needed to effectively manage their power grids.

The use of AI in both short- and long-term power market forecasts and scenario simulations will help utilities be ahead of the curve, when it comes to effectively managing the power grid in areas expecting a high penetration of DER assets. A key goal for AI-driven long-term power market simulations is for utilities to grasp a better understanding of customer energy consumption behavior over time and the likelihood of customers investing in a behind-the-meter generation. Without advanced, data-driven AI models that can accurately predict long-term supply and demand fundamentals, utilities could be at risk of having mediocre or subpar long-term power market forecasts which could lead to substantial negative financial consequences.

Meter to cash

There are many sub-segments in utility meter to cash efforts that can leverage AI, whether it is in areas such as debt collection, billing and metering or through customer interaction services. AI, if deployed properly in these areas, can be of major use and provide many benefits to utilities. Having AI models accurately predict customer payment behavior can help utilities address unpaid balances before they become a major issue. AI models designed to detect anomalies in customer payment or energy consumption behavior can ensure that billing and energy usage data are accurate before being shared with customers. Additionally, AI being deployed in customer interaction activities can help improve customer satisfaction and automate customer service functions when AI models can predict and address specific customer complaints or customer preferences in utility-run programs.

There is a wealth of data that can feed AI models from customer information systems (CISs). AI predictive models can be deployed in the meter to cash function within a utility and can provide accurate, actionable insights across finance, customer acquisition and retention, credit collections, while also predicting increased participation in utility programs such as demand side management and solar and electric vehicle offerings. Understanding and accurately forecasting utility customers’ interest and participation in clean energy programs will help utilities in reducing CO₂ emissions in their net-zero carbon emission initiatives.

Field services

In the areas of field services utility technicians can also benefit greatly from the use of AI. Predictive and prescriptive maintenance models can ensure utilities address asset issues and failures before they occur. AI models leveraging data from work order history can provide utility technicians with a wealth of insight on prior asset issues, equipment inventory and resolutions which in turn can support quicker and more efficient asset outage restoration times. The use of AI in asset maintenance and field work can reduce operational costs, improve asset downtime, increase technician productivity and also limit utility truck rolls on unnecessary regularly scheduled inspection rounds.

Fieldwork productivity gains and operational cost savings can be achieved by analyzing and understanding historical data on specific equipment and assets which have had operational, performance issues or mechanical problems in the past. AI models can provide utilities a path to a condition-based maintenance approach to operations as opposed to a traditional reactive or time-scheduled approach to operational maintenance which is less efficient and more expensive in the long run. Arming utility crews with condition-based actionable intelligence derived from AI models built off of historical asset, equipment and maintenance data can lower overall operational costs and also provide utilities with increased safety and productivity in the field which can also lead to longer asset and equipment life cycles.

Advice for power and utility companies investing in AI

• Before investing in AI, thoroughly research utility industry use cases that have proved to be successful and ones that provided positive business, financial and operational outcomes. Having a clear understanding of the outcomes and improvements desired, along with concrete target metrics to be achieved from the use of AI applications, will help secure funding and build support for key stakeholder buy-in.
   
• To avoid pitfalls when deploying AI, ensure there is access to all of the data needed to build robust AI models and make certain the data has been cleansed. AI models will only be as good as the data that is fed into them. Large volumes of data can be used in AI models, that being said, knowing where the data resides, the governance structure for the data to be used, and a thorough review of the accuracy of the data will be key in achieving positive results when using AI models.
   
• Be sure to not get stuck in endless pilot programs. When deploying AI models consider working with third-party AI vendors with strong power and utility industry domain expertise that can fill in skill gaps and assist your organization’s employees in building out AI tools. Create a definitive end date for pilot programs to get AI models to operate and provide results as soon as possible, in a live production environment. Taking the leap from pilot programs, to having AI tools in commercial production, will be a critical step in paving the way to reaping long-term benefits when putting a solid AI foundation and strategy in place.

AI applications in the power and utility sector are expected to grow and as use cases start to produce tangible results and demonstrate high returns on investment. AI technology and applications are not expected to replace human functions immediately in the power and utility sector. However, AI will certainly increase efficiencies and automation in many functions throughout power and utility organizations. AI applications, when implemented correctly, can provide the power and utilities sector with a wide range of benefits including increased efficiencies, productivity gains and ultimately better financial performance, which can impact power and utility companies' bottom lines.