Itemizing Your Energy Bill with Artificial Intelligence
The United States is the single most privileged country on the face of this planet by far, yet inexplicably some of her citizens believe that they’re somehow oppressed. As they bask in the safe glow of their temperature controlled homes, spewing froth vitriol over the Internet using their expensive laptops and smartphones, they fail to realize that the mere fact that they have electricity with close to a 99.99% availability is nothing short of a miracle. The massive amounts of infrastructure that were engineered and built to provide sustained electricity, the men that go out in the freezing cold and blazing sun to repair said infrastructure, and the efficiency and dependability at which the whole thing operates, is taken for granted by nearly everyone. At least once a month though, we’re all given a chance to complain about it when the energy bill arrives.
To be fair, your bog standard energy bill is still fairly archaic. It shows some dollar amount that usually goes up or down depending on the season, along with a rudimentary graph that shows your energy usage over time. It looks something like this:
If you’re looking to understand a bit more about how much energy you use in your household, that piece of paper isn’t very insightful. Wouldn’t it be great if you could receive an itemized bill which showed exactly what appliances consumed that energy that you’re paying for? That may soon become a reality thanks to the wonders of artificial intelligence.
We’ve talked before about how Google used their own artificial intelligence creation, Deepmind, to cut their power bills by up to 40%. Other companies have followed suite like Verdigris Technologies which helps commercial buildings cut their energy bills. Now we’re going to take a look at another company that just took in a pretty meaningful funding round last month in the pursuit of using machine learning to help you better understand your energy consumption.
If you’re feeling oppressed, try spending some time in India where the trains are so packed that people often get killed on their way to work at extremely competitive jobs with long hours and little reward for which they are extremely grateful for the opportunity. While you’re spending time in India, you may also come away with a newly found appreciation for dependable bidgley (that’s the Hindi word for electricity). It’s also the name of a Sunnyvalle, California startup which was founded in 2010 and has since taken in close to $52 million in funding to develop software that uses artificial intelligence to itemize household energy bills. So how does it work?
What Bidgely does is referred to as “energy disaggregation” and it involves itemizing consumers’ energy data into individual appliances using machine learning algorithms without any hardware needing to be installed. That’s right. Bidgely performs its disaggregation through meter data alone without the customer needing to install any hardware whatsoever. They claim to be the industry leader in “disaggregation technology” which works for both smart meters and monthly-read meters. If you’re unfamiliar with what a smart meter is, it simply allows for two-way communication between the meter and the power company. It’s these smart meters that allow Bidgely to do what they do.
Bidgely’s patented energy disaggregation technology detects and extracts appliance fingerprints from smart meter data. The data is converted into itemized energy bills and useful insights. Leveraging a database of over 50B meter readings from smart meters, Bidgely’s machine learning models extend itemization to non Smart Meter homes using a “matched peer” or a “matched region” approach.
That’s pretty remarkable that they can perform actual itemization without installing any hardware. It’s not just statistical models that provide “typical usage” either. Bidgely says that if the consumer goes on vacation or upgrades an appliance, the itemization will reflect the change. One of the examples they provide is when a utility company offers a new air conditioning unit and claims that it will offer energy savings. Using a statistical model, energy consumption will be reduced, but the percentage attributed to “cooling” will not change. Using proper “energy disaggregation” will show the change so the customer can actually see the impact:
They claim to offer this solution at a fraction of the cost that hardware-based solutions would incur and they can reach 100% of homes out there, regardless of the types of meters being used. Well, only those homes that are lucky enough to have electricity in the first place of course, which is a market of 5.6 billion people at the moment. At the moment they’ve only captured 10 million homes, thus the need for their latest funding round to scale the solution as quick as possible. They’re working in 10 countries at the moment, which shows that borders won’t present any barriers to their future expansion plans.
The platform is cloud-based which means that utilities can simply plug in to it and begin offering the solution under their own brand name. In addition to telling people which appliances are consuming the energy they’re paying for, there are also other ways to engage customers such as through gamification. The proof is in the pudding as they say, and a look at their customer list shows some big name utilities have signed on:
Two of the energy names seen above are also investors in Bidgely. German company E.ON is one of the world’s largest investor-owned electric utility service providers with more than 56,000 employees and 2015 revenues of $144 billion. The second large utility invested in Bidgely is Exelon (NYSE:EXC), a publicly traded utility that generates revenues of approximately $34.5 billion and employs approximately 34,000 people. Utilities have more interest in just conserving energy since all that big data that Bidgely is capturing will be useful to aggregate and analyze for additional insights. That’s one of the goals that Bidgely has alongside their latest funding round – “business intelligence applied to broader decision-making within utilities and energy retailers“. It’s only a matter of time before we all have itemized electricity bills arriving in our inboxes thanks to the wonders of artificial intelligence. Well, those of us that are fortunate enough to have electricity that is.
Here at Nanalyze, we invest the majority of our savings in 30 dividend-paying stocks that increase our income every year and outperform during both bull and bear markets. Find out which ones in the Quantigence report freely available to Nanalyze subscribers.