Funding has been allocated to support a variety of assistance programs, from direct bill subsidies like the Low Income Home Energy Assistance Program (LIHEAP) and arrearage management plans to weatherization rebates and time-of-use rate restructuring. But, according to the American Council for an Energy-Efficient Economy, only 13 percent of low-income households are on the receiving end of support.

Closing the gap between aid eligibility and actual enrollment means prioritizing technology that enables electric companies to identify and engage all income-qualified customers. By leveraging AI-driven analytics, electric companies can move beyond broad census data, transforming consumption patterns and appliance usage into clear indicators of energy vulnerability across an entire customer base.

Specifically, AI decodes the unique “fingerprint” of every home, allowing electric companies to not only detect customers with high energy burdens at the individual household level but also accurately match customers with relevant programs and messaging.

Identify Inefficient Applications

Aging heating and cooling systems and failing water heaters are the two biggest drivers of energy burden, making these homes ideal for targeted equipment replacement and electrification subsidies.

Distinguish Between Structural Inefficiencies and Behavioral Spikes

Homes with poor insulation are best served by weatherization upgrades, whereas homes with dramatic spikes are prime candidates for demand response programs.

Streamline Enrollment Via Behavioral Proxies

Because specific consumption signatures correlate with socioeconomic need, electric companies can auto-enroll customers in assistance programs and remove the administrative paperwork that often prevents households from receiving aid.

Identifying an income-qualified household is only half the battle; the other half is earning the trust required to help them. AI enables a shift from mass marketing to personalized coaching, so that rather than sending generic energy-saving tips that feel out of touch with a struggling family’s reality, electric companies can provide specific, timely insights. 

For example, a notification shouldn’t just say “save energy,” it should alert a customer that their HVAC unit is underperforming and provide a direct path to a pre-qualified repair grant.

Layering generative AI (GenAI) onto these data insights, electric companies can even tailor the timing, language, channel, and tone of individual customer outreach to ensure assistance offers arrive exactly when they are most relevant—such as immediately following a high-usage weather event. Because when communication is rooted in a customer’s actual consumption reality, the likelihood of both participation and program 
success increases.

The whitepaper, “Exceeding Affordability Goals With UtilityAI: Leveraging AI and Behind-the-Meter Data to Better Serve Income-Qualified Customers,” details how these data-driven strategies enable precise, high-impact support for those who need it most.

Electric company leaders are recognizing the only sustainable approach is to drive efficiency, extract value, and re-invest in existing investments. Thankfully, tools like software and Al present solutions to meet these needs.

The Power of Layered Technology 

The secret to maximizing Al's value for electricity companies lies in combining two types of technology: 

• Horizontal Al
  • This is the broad, general-purpose technology, like global cloud services, big data platforms, and large language models from Google, OpenAI, Microsoft, and Amazon, that provides foundational processing power and tools used across all industries.
• Vertical Al
  • These are highly specialized models built for specific industries. In the energy industry, this includes machine learning algorithms with expert knowledge on smart meter usage patterns to identify appliance and DER signatures, derive relevant consumer insights, and predict risks like wildfires or transformer failures. 

By combining the strength of the general platforms with the specialized knowledge of vertical models, electric companies can enable industry-specific use cases that unlock grid flexibility, support distribution planning, and manage costs for customers through optimized rate design. 

Achieving Real Returns from Al Investments 

Successful Al initiatives start by defining a clear business goal, not by chasing a new technology. For example, to improve grid stability, grid planners need to know more than just total energy use. They need granular data that identifies the driving factors behind changes in load shape.

When Al is able to identify five EVs charging simultaneously on the same feeder, at the same time summertime air conditioning use spikes, grid planners can now target customers in these neighborhoods for incentivized load shifting programs.

Achieving measurable, scalable goals that clearly impact load, consumers, and the grid will drive wider adoption of new technology and create a virtuous cycle where efficiency lowers costs, maintains affordable rates, and generates public support. 

Deploying Al Safely in the Energy Industry 

To scale responsibly, innovation cannot be a "black box" that operates in isolation. With large electric companies often hesitant to send sensitive customer data, like AMI information, outside their secure environment, the solution is to deploy the advanced vertical Al models directly inside the electric company's own environment.

This strategy achieves two things:

• Security And Flexibility: Data remains secure and ensures the specialized tools work seamlessly with the electric company's existing investments in general platforms. 
• Empowerment: Specialized models turn into an essential data asset that the electric company's own data scientists and engineers can use. This empowers them to quickly build new applications, rather than relying on an external vendor for every single use case.

Bridging the gap between broad digital infrastructure and energy-specific intelligence allows electric companies to maximize their existing data ecosystems, turning years of accumulated AMI data into a high-yield strategic asset.

Given the complexity and rapid evolution of AI, the Scaling AI in the Energy Industry episode of the Electric Perspectives podcast, featuring Arizona Public Service's (APS's) Michelle Ferrara, offers a deeper exploration into how APS is building an AI ecosystem to strategically support their technology investments. Learn more at bidgely.com/utilityai-pro.

When you consider the next level of analysis, as 10 or 20 residential properties roll up to a single transformer, the importance of understanding the individual demand characteristics of each home becomes even greater. A transformer's load shape is the aggregate of the load shapes of the homes that it serves. And, as those 10 or 20 homes on a single transformer increasingly electrify, they can create rapid, unexpected demand spikes on that transformer.

Leading energy companies are now leveraging behind-the-meter intelligence to understand each customer's energy usage patterns and rolling up those granular insights to provide comprehensive views of transformers, feeders, and substations. This bottom-up view enables more informed decisions about what infrastructure improvements are needed today in order to maintain grid reliability and resilience.

Even more compelling, behind-the-meter intelligence can inform scenario planning that enables data-driven decision making long into the future. By modeling various policy, consumer behavior, and economic scenarios—such as different rates of EV adoption or building electrification—energy companies can make informed decisions about both traditional infrastructure investments and non-wire alternatives. Generative artificial intelligence is now amplifying this scenario planning capability by simulating multiple permutations and combinations of input parameters, enabling grid planners to evaluate countless nonwire alternative strategies and predict individual asset performance across different timeframes.

At Bidgely, we are working with a wide range of energy companies to implement this next-generation grid planning approach. The urgency of this transformation cannot be overstated. While many energy companies may not yet have reached the electrification or distributed energy resource adoption tipping points, those tipping points could arrive within just one to two years. Energy companies that begin implementing comprehensive scenario planning will be positioned to successfully manage the energy revolution now underway.

The shift from viewing customers as uniform loads to understanding them as unique energy participants represents a fundamental change in energy company planning. Those embracing this data-driven, customer-centric approach will build more resilient, efficient infrastructure capable of supporting new supply and demand paradigms while maintaining the reliability customers expect.

Amazon's rise to a $2.5-trillion market cap exemplifies this concept perfectly. As customer numbers grew on their platform, more sellers and products were attracted to join, which in turn drew even more customers—creating an ever-accelerating cycle of value and growth.

Today, the electric power industry stands at a similar inflection point. Unlike the traditional linear approach to industry operations, the modern grid is evolving into an integrated ecosystem as electric vehicles (EVs), renewable energy sources, and battery storage systems enable customers to actively participate in grid management.

This industry flywheel begins with customer engagement. As distributed energy resources gain widespread adoption, there are significant variations in both geographical distribution and device types. Treating customers regionwide as a uniform cohort no longer works. Companies must employ hyper-personalized messaging tailored to each customer's unique appliance usage, EV/solar asset ownership, and lifestyle preferences.

By leveraging artificial intelligence (AI), companies can detect who has EVs, solar panels, or inefficient appliances, as well as define energy use patterns for each. This sets the flywheel in motion: delivering exceptional, personalized experiences that foster positive customer relationships and motivate active grid support participation.

The second part of the flywheel focuses on grid planning. Behind-the-meter customer energy use data provides grid analysts with critical insights into where assets face constraints. AI-enabled platforms can identify which customers own specific devices and their usage patterns, and companies can pinpoint those with the greatest load shaping or shifting potential to alleviate grid constraints.

This leads to the final component: load management. Among the various approaches to manage peak load, load shifting and load shaping prove most effective for addressing long-term grid stability as distributed energy resources scale. With AI-based disaggregation of behind-the-meter data, companies gain access to energy usage by appliance type and consumption patterns, enabling precisely targeted program recruitment.

Because companies have already delivered exceptional, personalized experiences, target customers are primed to become willing participants in demand response programs. This reduces overall energy costs for customers, enhancing their experience and bringing us full circle to the start of the flywheel, where it continues to turn with increasing momentum.

Just as Amazon's flywheel inspired a retail evolution, the energy space is now poised for its own transformative f lywheel transition. Please reach out to continue the conversation.