Blend launches predictive engine for mortgage automation

The company turns to machine learning for Blend Intelligence

Blend launches predictive engine for mortgage automation

Blend, a digital platform provider for consumer lending, has introduced Blend Intelligence, a predictive engine aimed at automating the mortgage experience to make it more efficient for lenders.

The company turned to machine learning to develop the functionality. Blend said it is designed to learn from how lenders interact with the Blend platform to continuously improve its performance over time.

According to the company, Blend Intelligence lowers lender processing costs while at the same time giving a more real-time experience to consumers. The functionality relies on a novel approach to the mortgage origination process.

“Let’s take Blend Intelligence’s predictive modeling as an example of real change that technology can have on loan origination,” wrote Grace Qi, product manager for Blend Intelligence, in a blog post. “It may be easiest to think about the end result as a powerful recommendation engine. In the same way that Netflix recommends movies based on your past viewing behavior, Blend Intelligence recommends loan conditions based on past lender behavior.”

Blend Intelligence would automatically request pay stubs and other common documents, following simple rules. However, the recommendation engine can also suggest follow-up requests for rarer conditions like terms of withdrawal or a profit and loss statement.

Other key features include loan health checks that happen in the background to automatically spot and flag bad data and automated suggestions to follow-up the borrower and third parties to shave more days off the process.

Blend said loan officers and underwriters can use the functionality as an important safety net because predictive modeling can help figure out what conditions are missing upstream rather than downstream, especially those conditions that aren’t immediately obvious.

“We use this added layer of machine learning because we know that basic rules aren’t enough for the complex loan industry. We also know that this technology will help lenders achieve lower costs and more efficient loan processing because there is less need for loan team intervention. This is where the magic really happens,” Qi said.