How AI can improve utility storm response, with increased reliability, lower ratepayer costs

The frequency of extreme weather events is increasing, with Gallup insights showing over a third of U.S. adults have been affected by them in the past two years. Not only do these events cause devastating losses to communities in terms of death and displacement, but they’re also taking a massive economic toll. Since 1980, severe storms have resulted in a whopping $455 billion in damage, with a third of these happening in the past five years alone.

Yet damage control isn’t the only problem to contend with. Traditional methods of storm response come with their own hidden costs that are putting a significant strain on the grid, impacting its resiliency and reliability, and making affordability a major issue for ratepayers.

Here, AI can make storm response more efficient and effective for all stakeholders — utilities, ratepayers, vendors, regulators and the various governments involved. Modernizing storm response using AI will boost grid resiliency and reliability while directly impacting ratepayers' affordability. A careful, balanced integration of AI alongside existing traditional storm response methods is the way to go in the future.

Storm recovery can be eye-wateringly expensive, and utilities are often forced to pass on these costs to customers, which can be a burden for years. In 2022, utilities faced $12.4 billion of weather-related debt, which American utility customers had to help cover. Recently, Duke Energy requested authorization from Florida regulators to pass on $1.1 billion to customers for recovery costs.

For starters, quickly mobilizing qualified crews and ensuring personnel — assessors, wire-down guards and lineworkers — are available for swift restoration demands massive coordination and manual effort. Currently, chaos and clutter are the name of the game.

Crews also encounter extremely dangerous conditions like downed wires, felled trees and flooded roads, causing significant safety hazards for responders. They’re racing against the clock to make informed decisions, such as mapping out blocked roads or areas of flooding. Delays and complications in communicating these safety hazards not only cause further safety threats but also ramp up restoration costs.

Another hurdle for stretched resources is accurately tracking time and expenses, which is integral to justified cost recovery. In a storm’s aftermath, utility companies often have to chase paper trails and missing information, preventing cost recovery from regulators.

For vendors, this can lead to significant cash flow issues, given that millions of dollars of unpaid and delayed invoices end up sitting in limbo for months on end. Privately owned brokers are frequently involved in organizing external crews, facilitating time and expense tracking, and even paying contractors directly to provide short-term liquidity. However, they can add tens or even hundreds of millions of dollars to storm recovery costs, which often end up on utility customers’ shoulders.

And the toll of these challenges is only hitting utilities harder. Last year, the U.S. faced its 14th consecutive year with over 10 separate billion-dollar weather events. Meanwhile, per a 2024 rate disallowance case study by KPMG, disallowance metrics have increased from 10-20% between 2008 and 2018 to 35-40% between 2019 and 2023, meaning utilities are also under more pressure and scrutiny to justify capital investments to regulators.

Imagine an Uber-like solution for storm response and restoration. In the age of AI, utilities can track and connect resources to predicted areas of storm impact. Combined with AI-powered asset inspection and vegetation management, this approach makes storm response significantly more effective.

How can this be done? AI platforms can forecast high-impact storm areas, and some U.S. utilities are predicting fuse failures during storms due to vegetation issues with more than 80% accuracy. Combining that information with real-time qualified resource availability and skills matching helps with the storm response and restoration efforts, leading to better SAIDI, SAIFI, and CEMI reliability metrics. It’s even possible to proactively perform last-minute vegetation management in the area of predicted fuse failures to lower the impact of outages during a storm.

AI-powered platforms can also slash decision-making timeframes by bridging the gap between field resources and office workers, ensuring all teams are able to collaborate and make informed decisions collectively. This collaboration speeds up infrastructure damage evaluations and helps keep a pulse on restoration progress, critical for teams that are stretched thin.

Automated timesheets are a game-changer for tracking time and expenses, digitally capturing location and task details of each crew, plus receipts. This improves transparency for up-front cost verification and reinforces regulatory compliance while potentially reducing disallowance metrics for utilities. Vendors also benefit since accurate invoices and documentation of completed work speed up payment processes, strengthening cash flow so vendors can continue running services for utilities. This can reduce the need for private brokers and save the utilities and their ratepayers hundreds of millions of dollars a year in storm recovery costs.

Now more than ever, with the advent of AI, utilities need to turn to more cost-effective strategies that carefully integrate technology into storm response and restoration. The capital saved could be invested in other areas, like proactively strengthening the grid against future disasters via grid-hardening projects.

AI shouldn’t replace humans but rather enhance their capabilities in severe storm response while providing more opportunities for transmission and distribution resources.

This approach also alleviates doubts that stall AI adoption. Many utility workers are wary of automation, given the widespread perception that it’s here to take over their jobs. The reality is that in storm response, automated platforms can create opportunities for resources in restoration efforts that might not have existed before. To tackle AI misconceptions, people need to be trained and familiarized with how tools empower their efficiency and decision-making when disaster strikes.

Moreover, AI on its own is not foolproof, and human involvement cannot be removed from any scenario. It’s crucial for storm response teams to understand how they can best leverage AI for greater and safer efficiency in storm response. Investing in hands-on AI training is a key avenue to building resilience and preparedness for future storm events.

Rather than jumping at the first AI solution they encounter, utility companies must carefully evaluate their existing processes and identify gaps and opportunities for improvement. This enables them to not only invest in the right tools for their specific needs but also to know when and where these can make the most impact for stronger, faster and more efficient storm response.