How AI & machine learning shapes energy providers’ COVID-19 response

By Siddhartha Sachdeva

The impact of COVID-19 took everybody by surprise, including energy suppliers. Within a couple of weeks of March 2020’s shelter-in-place orders, Innowatts was able to predict deviations in standard consumption patterns due to unexpected changes in consumer behavior. For many the home became – and still is – a central hub for education, entertainment, and business. Accordingly, domestic energy usage rose at an unprecedented rate.

Our analysis – as featured in our customer NRG’s 2020 Sustainability Report – revealed that residential electricity consumption rose by approximately 8 percent on average during Q2 2020 and was up to 15 percent higher during working hours. Conversely, commercial energy usage slumped by as much as 30 percent.

During the extreme weather events in Texas which caused rolling blackouts and power shortages, Innowatts applied its AI-driven algorithms to meter-level data, in near real-time, identifying customers that had been incorrectly categorised as gas heating only, when electric heating was their principal source.  Innowatts then was able to predict and map customers in southern Texas that would drive a surge in electricity demand as the temperature dropped and those customers who deployed temporary heaters to stay warm. Energy suppliers then were able to re-profile their customers, creating a much more accurate forecast and subsequent procurement strategy for the day ahead.

A more accurate day-ahead forecast vastly improves the energy supplier’s ability to hedge their market positions thereby reducing pricing exposure on the spot market – where wholesale electricity prices soared as high as $9,500 per MWh. Understanding individual demand drivers thus enables effective pricing for customers and optimizes their energy usage during periods of high demand and systemic stress.

Today, the US economy has rebounded from the worst of the downturn and Innowatts’ AI forecasts from 40+ million meters have enabled us to see exactly which parts of the economy have recovered faster than others. For example, power consumption recovery for some business segments such as banks, restaurants, fitness centers and furniture stores were slower than that from convenience stores, gas stations and grocery stores.

With greater predictive, actionable and operational and customer intelligence, Innowatts platform users have been able to develop more accurate predictions to account for the shift in commercial and residential energy use. These insights will be just as valuable as the economy continues to rebound. As we continue to move through unchartered territory, real-time, AI-driven interval data analysis gets us as close to forecasting the unprecedented as possible, returning power into the hands of the beholder. With it, energy suppliers and utilities can continue to support their residential, small business, commercial, and industrial customers on the journey to recovery in the most proactive and sustainable way.

To find out more about how AI-driven insights could help you better engage with customers, get in touch.

Join our mailing list to to get notified of
new case studies, white papers, news and events.