Turning customer data into grid intelligence [partner content]

Episode Summary

Executive Summary: The episode explores how PSEG Long Island and Bidgely use AMI data and AI disaggregation to support time-of-day rates, improve affordability, and manage load growth on a weather-exposed, mostly residential grid. The discussion shows how appliance-level insights, targeted customer engagement, and trust-based utility-customer communication can reduce peak demand and delay costly grid investments.

Main Topics: Long Island’s grid and weather-driven challenges (Priority: 5/5): Lou DeBrino explains that Long Island’s narrow, coastal geography makes it especially vulnerable to hurricanes, snow, and hot summers, increasing pressure on a mostly residential utility system. Time-of-day rates and affordability (Priority: 5/5): PSEG Long Island rolled out variable pricing to better reflect real energy costs and help customers shift usage away from expensive peak periods, while maintaining affordability. Bidgely’s appliance-level disaggregation (Priority: 5/5): Bidgely turns meter data into appliance-level insights, letting customers and utility staff see what drives usage and bills, analogous to turning a vague bill into a detailed credit card statement. AMI evolution and real-time customer intervention (Priority: 4/5): The discussion traces the move from monthly reads to 15-minute AMI data, and then to AMI 2.0 where analytics can run on the meter for real-time alerts and customer actions. Targeted programs for EVs, HVAC, solar, and efficiency (Priority: 5/5): Using disaggregated data, the utility can identify EV owners, solar customers, air-conditioner and pool-pump users, and weatherization needs to deliver more precise efficiency and demand-response programs. Non-wires alternatives and grid planning (Priority: 4/5): The speakers explain how behind-the-meter visibility helps distinguish load drivers, assess what demand can actually be shifted, and determine whether wires, batteries, or demand-side management is the right solution. Trust, personalization, and the utility-customer relationship (Priority: 5/5): Both speakers emphasize that effective programs depend on customer trust, transparency, and highly personalized communication, increasingly enabled by AI and digital channels.

Key Arguments: Time-of-day rates are more effective when customers understand the specific appliances and behaviors driving their bills, not just the total amount owed. Machine-learning disaggregation converts raw AMI data into actionable appliance-level intelligence, enabling more accurate customer guidance and utility targeting. 15-minute AMI data significantly improves over old monthly reads and supports scalable, territory-wide customer programs without physical installs. AMI 2.0 could allow real-time, on-meter analytics and push alerts, creating immediate interventions when customers are using energy at peak times. Detailed usage data helps call center agents become energy advisors, improving customer service for high-bill questions and program enrollment. Targeted outreach is more efficient than mass marketing because it focuses utility spend on customers most likely to benefit from specific programs. Behind-the-meter visibility can inform grid investment decisions by showing whether stress is caused by flexible loads like EVs or less-flexible loads like HVAC. Trust is essential: utilities must prove that data is accurate, useful, and deployed in the customer’s interest to sustain long-term adoption.

Data Points: Eligible residential customers on time-of-day rates: 900,000+ - PSEG Long Island moved all eligible residential customers onto variable pricing Utilities served by Bidgely: 45+ global utilities and energy providers - Bidgely’s platform footprint Patents in machine learning disaggregation: close to two dozen - Bidgely’s intellectual property base for appliance detection AMI read interval: 15-minute increments - PSEG Long Island collects consumption data every quarter hour after AMI rollout AMI deployment completion: end of 2020 - PSEG Long Island completed residential AMI rollout by this time Meters in AMI 2.0 work: close to 2 million meters - Bidgely is working with major meter manufacturers on live on-meter analytics Episode count in series: second in a four-part series - This podcast is part of a broader series on AI in utilities First episode publish date: June 16 - Referenced in the outro as the prior episode in the series

Pivotal Quotes: "We have close to two dozen different patents in what we would describe as machine learning disaggregation" — Ted Nielsen: Explaining Bidgely’s core technology for identifying appliance-level usage from meter data "It turns it into a credit card statement so that you can see each of the charges and each of the appliances that are actually driving that $100 bill." — Ted Nielsen: Using an analogy to describe how disaggregated data helps customers understand their energy use "The first one is: it's very hard to intervene strategically if you don't actually understand the problems." — Ted Nielsen: Summarizing the need for appliance-level insight before designing grid interventions

Implications: Utilities can better manage load growth and affordability when they pair AMI data with AI-driven disaggregation and personalized outreach. The future grid will depend less on blunt infrastructure expansion and more on trust-based, data-informed customer participation.