Steven Heruty, our Chief Credit Officer, is here to talk about MosaicScore and how it can help more customers get approved for solar loans and helps Mosaic continue to deliver high-quality assets.
Question: "Thanks for taking time out of your busy schedule to talk to us about MosaicScore! Firstly, can you tell us more about your role and responsibilities at Mosaic — and what made you want to work here?"
Steve: "Thanks for taking the time to speak with me. In my role as Chief Credit Officer, I manage the credit quality of the loans originated through the Mosaic platform. I also manage our talented Data Analytics team which helps our internal customers answer important business questions with data. I joined Mosaic to support the transition to clean energy that residential solar helps to drive. Only 7% of homes in the US have rooftop solar, so we have a big opportunity for growth."
Question: Can you give us a brief description of what MosaicScore is and how Mosaic’s technology teams have worked on a new and improved version of the feature?
Steve: "The MosaicScore is a custom-built proprietary risk model used in the underwriting decision, or the decision to approve or decline during prequalification. Most people are familiar with the FICO score which is a generic risk score widely used by lenders. Experienced lenders will typically build a custom risk model based on their own customer data to complement their use of generic risk models. A custom model is more finely-tuned to a specific lender's applicants and can provide incremental predictive power over a generic model built on the general population and particularly exciting within residential solar lending. In other terms, MosaicScore helps us look beyond credit scores to approve homeowners.
This latest version of the MosaicScore has been completely redeveloped using more recent data, a broader set of credit attributes and advanced machine learning techniques that bring all this information into a predictive score for each applicant."
Question: How does machine learning help analyze the data we have? Is our machine learning the same as AI?
Steve: "Machine learning is a subset of AI that has been around for decades and includes techniques for using computers to analyze data to identify relationships that can be used to predict outcomes. MosaicScore was developed with a form of supervised machine learning used to build an optimal set of decision trees using dozens of credit attributes and recently updated with new data that weren't available in the prior model. The broader category of AI also includes unsupervised machine learning techniques used in LLMs which focus on discovering patterns and structures within unlabeled data. MosaicScore uses supervised machine learning to identify and optimize the use of credit attributes into a model, so professionals and homeowners can benefit from real-time decisions."
Question: Can you explain the customer credit attributes and how they help more homeowners get financing?
Steve: "A credit attribute is a piece of information about your credit history. This could include how long you have used different types of credit, how much credit you have been given, and how well you have managed it, or how often and recently you have applied for new credit. In general, the longer you have responsibly used different types of credit, the greater your chances of being approved for financing."
Question: How does MosaicScore benefit clean energy professionals, especially sales representatives?
Steve: "The MosaicScore is an innovative, refined tool specifically developed for residential solar; compared to generic qualification criteria used by other solar lenders which can exclude too many good consumers. Our new proprietary model has also been strengthened with updated data for a more customized credit decision experience. This helps us achieve highly competitive approval rates while maintaining loan quality. And higher approval rates can translate to more completed sales for professionals and more homeowners benefiting from energy independence and utility bill savings."
Question: What excites you most about MosaicScore?
Steve: "The development of the new MosaicScore leveraged an advanced machine learning technique that reviewed thousands of possible credit attributes to determine the best combination of factors that resulted in the final model. In addition to being developed from our long history of customers, the model also considers performance data on solar loans financed through other lenders."
Question: Is there anything else you want to share?
Steve: "Our model underwent a thorough evaluation with an extensive testing period and complex audit. Our dedicated teams worked diligently to make this new model a reality and improve the installer and homeowner financing experience. We are excited to share the launch with our partners and continue to innovate to provide best-in-class services!"
Financing applied for and processed through the Mosaic platform is originated by Solar Mosaic LLC or one of its lending/financing partners. Equal Housing Lender