with Buzz Solutions co-founders Kaitlyn Albertoi and Vik Chaudhry, they shared their backstories in California and New Delhi, talked about the pivots in their startup, and how their transmission and distribution imaging platform is saving utilities 50% of time and effort for analysis.
Identification of an issue in a component is challenging. The solution has to be able to identify the asset and what’s wrong with it. In machine learning, reinforcement learning makes image recognition better. Buzz Solutions trains its neural net with base data from utilities and provides that basic model to their customers. But then customers have a human-in-the-loop model to improve their instantiation of the model. If an insulator is marked by the AI as damaged, a human can override that.
What Buzz Solutions is seeing in the utility sector is that linespeople and engineers, once they’ve trained the model, are able to do their primary job, repairing lines and components or doing resiliency engineering. The models and algorithms just help them find the things to fix. Buzz Solutions assists utilities with grid resilience, reduces increases in labor costs, increases speed of identification of critical issues, and enhances employee satisfaction. Making workers happier is a counter-example for a lot of automation solutions. Buzz Solutions surveys workers at the end of the first week, first month and a few months down the road, and sees satisfaction dramatically improves as they see how much time is being saved and how their work days are improving.