How Smart Meters and AI Uncovered 37,000 Hidden Families in Need

LOS ALTOS, Calif. – October 26, 2026 – As a former market analyst, I spent years dissecting utility earnings reports, looking for margin improvements, peak load optimizations, and infrastructure investments. It is easy to get lost in the numbers and forget the end consumer. But behind the dense data tables of a recent corporate announcement from enterprise energy software provider Bidgely and utility giant Pacific Power, a very human story emerged. The two companies recently secured the 2026 Association of Edison Illuminating Companies (AEIC) Achievement Award, a prestigious peer-reviewed honor in the utility sector. The recognition wasn't for a new power plant or a breakthrough in battery storage. It was for utilizing artificial intelligence to find 37,000 vulnerable households hiding in plain sight and connecting them to critical financial relief.

The initiative represents a fascinating pivot in how electric utilities operate. For years, the industry has aggressively deployed smart meters primarily for automated billing and grid management. Now, companies are repurposing those same data streams to combat energy poverty, raising compelling questions about the intersection of machine learning, social welfare, and consumer privacy.

Beyond the Plateau: Data for Good

Following the economic turbulence of the COVID-19 pandemic, Pacific Power established the Oregon Low-Income Discount Program. The initiative was designed to deliver vital bill discounts to income-qualified households, a mandate closely monitored by the Oregon Public Utility Commission (OPUC). Initially, the program saw steady enrollment. However, participation eventually hit a brick wall. Automatic enrollment, self-attestation, and broad demographic targeting reached their absolute limits, resulting in a frustrating two-year plateau.

The utility knew more families needed help but could not easily identify them. "Traditional outreach methods only go so far, and we knew there were more families who needed support but fell through the cracks of conventional data searches," said James Owen, vice president of customer experience and innovation at PacifiCorp, Pacific Power's parent company.

To break the stalemate, the utility deployed Bidgely’s UtilityAI Analytics Workbench. The platform allowed the company to pivot from generalized marketing to hyper-targeted outreach. The results were immediate and measurable. The program broke its two-year plateau, achieving a 2.3 percent conversion rate in its first pilot wave and a 1.7 percent conversion rate in the second. Overall awareness of the assistance program surged by 10 percent quarter-over-quarter, hitting 65 percent in an independent J.D. Power study.

"Working side-by-side with Bidgely allowed us to not only find these eligible but unenrolled households and connect them with critical bill assistance, but also build a framework that will serve our communities for years to come," Owen added.

The efficiency gains extended to the marketing budget itself. By utilizing AI to tailor outreach—including bilingual messaging that achieved a 28 percent open rate—the utility drastically reduced its reliance on wasteful direct mail. Print volume plummeted from 16,000 pieces to just 1,600, while response quality actually improved. Furthermore, participating households maintained a 90 percent customer satisfaction rating and accelerated their energy efficiency savings by 32 percent, proving that financial assistance and demand-side energy management can work hand in hand.

The Mechanics of Algorithmic Empathy

From an analytical perspective, the mechanics of this achievement are remarkable. How exactly does a software platform deduce that a family is struggling to pay its bills without accessing their bank accounts or tax returns?

The answer lies in a technology known as Non-Intrusive Load Monitoring (NILM). Bidgely’s platform analyzes high-frequency smart meter data to identify the distinct electrical signatures of individual appliances without requiring any additional hardware in the home. It can mathematically separate a household's total electricity consumption into specific usage categories, distinguishing the hum of a refrigerator from the draw of an HVAC system or a water heater.

To find candidates for the Low-Income Discount Program, the software combined census income data with these appliance-level usage insights. The algorithms looked for high-consumption households situated below specific income thresholds in census blocks, as well as customers on the cusp of qualifying who likely did not realize they were eligible. For instance, the system might identify consumption profiles typical of smaller homes, older and less efficient appliances, or behavioral patterns indicating a strenuous effort to conserve energy.

This represents a monumental shift from traditional utility operations. Instead of waiting for a customer to miss a payment or call a support center in distress, the utility uses algorithmic models to proactively predict vulnerability and offer a lifeline.

The Privacy-Relief Paradox

While the social benefits of this targeted outreach are undeniable, the methodology introduces a complex ethical tension. Combining granular energy data with census demographic records creates highly sophisticated behavioral profiles, sparking what industry insiders call the privacy-relief paradox.

Both Oregon and California maintain robust data privacy frameworks. The OPUC enforces strict rules restricting how utilities can collect, use, and share customer energy usage data, typically requiring consent for third-party sharing. Similarly, the California Consumer Privacy Act (CCPA) grants consumers sweeping rights over their personal information.

Consumer privacy advocates have consistently voiced concerns about the potential for smart meter data disaggregation to infer highly personal information. An algorithm that knows when your appliances are running can theoretically infer household occupancy patterns, lifestyle habits, and even the use of specialized medical equipment.

"When an algorithm can infer your household size or whether you are running an inefficient space heater to survive the winter, we cross a line from basic utility billing into behavioral profiling," noted one digital privacy advocate who monitors utility data practices. "The fact that this data is being used for social good right now doesn't negate the need for strict, transparent guardrails. If a utility can identify vulnerability, that same data must be fiercely protected so it cannot be exploited by predatory lenders or unauthorized third parties."

Bidgely's approach mitigates some of these risks by utilizing aggregated and anonymized patterns. The insights used for targeting are derived at a demographic segment level rather than explicitly linking individual smart meter data to specific personal income records. Still, as the technology becomes more precise, regulatory bodies will face increasing pressure to ensure that algorithmic welfare does not inadvertently become a tool for surveillance.

A New Benchmark for Utility Innovation

Despite the regulatory tightrope, the partnership between Pacific Power and its software provider illustrates the immense, untapped potential of grid modernization. The two organizations initially collaborated on basic residential Home Energy Reports and behavioral efficiency programs. Today, they are applying sophisticated data intelligence to solve some of the industry's hardest operational and societal problems.

"Pacific Power continues to set the benchmark for how forward-thinking utilities can apply AI beyond energy efficiency to solve critical affordability and equity challenges," said Gautam Aggarwal, President and Chief Revenue Officer of Bidgely. "We are thrilled to congratulate Pacific Power on this well-deserved recognition, and it is an honor to partner with a utility that prioritizes data intelligence to deliver meaningful value to the communities it serves."

The AEIC Achievement Award serves as a powerful validation of this approach. It signals to the broader market that the future of the utility sector relies not just on generating power, but on generating actionable empathy from raw data. As energy affordability continues to be a critical crisis across the nation, the ability to query, segment, and target individual customer needs with extreme precision will transition from an innovative luxury to an absolute regulatory necessity.