Artificial Intelligence (AI) is used often in the solar industry for its predictive capacity—for weather forecasting and to support informed decision-making. AI’s algorithms are able to analyze massive quantities of data from satellites, weather stations, and sensors to determine patterns and foresee weather events that could impact energy generation. We recently reached out to Attila Toth, founder and CEO of PowerScout, an AI-driven technology marketplace, to learn about other applications of AI in the solar industry. He offered insights regarding his company’s remarkable AI chatbot, a tool that helps users navigate solar investment decisions.
Distributed Energy (DE): How does the technology work?
Attila Toth (AT): The PowerScout Messenger chatbot gives Facebook Messenger users a chat interface where they can get a solar energy savings estimate for their home in seconds. The chatbot asks users their address in order to assess their home for solar panels. It will then tell them how big of solar panel system they might need, as well as how much they could save per month/year by switching to solar.
Behind the scenes, PowerScout uses machine learning algorithms and Artificial Intelligence-enabled computer vision models to assess the homeowner’s roof from high-resolution satellite imagery. It establishes how much solar energy a roof could generate and how much it would need to generate to meet the homeowner’s power needs. This figure is then combined with local electricity price data to figure out the potential savings from going solar.
Rather than having to leave Facebook or use an external calculator, the homeowner can get their solar installation quote in roughly 2 minutes from within Facebook Messenger.
DE: Why did PowerScout create the Chatbot?
AT: The average American spends nearly an hour each day on Facebook. We wanted to meet people where they are and allow them to get reliable solar estimates while using the Messenger Platform.
DE: How accurate are the estimates that the Chatbot provides?
AT: We’ve worked with hundreds of installers and thousands over the years and have consistently found our estimates are spot-on, down to the last panel and number of kilowatts a house will require. Furthermore, our machine learning models are constantly refining, testing, and improving themselves.
DE: Going forward, will it also be applicable for industrial inquiries?
AT: Definitely. We plan to roll it out for commercial and industrial applications so that a building owner can get this same easy experience and power their business operations on solar.