GPT is already disrupting the mortgage industry, says Figure, the former Brex COO who now heads the unicorn fintech

Lending startup Figure today announced the launch of AI tools to make the home loan process more efficient. The company will launch a GPT-4-powered AI tool to help catch errors in lending documents.

Founded in 2018, Figure specializes in helping consumers secure mortgage lines. The company claims its fully online process condenses what would normally be a 45-day process into five. More than half of Figure’s business is B2B, with companies like solar panel lending company GoodLeap.

According to PitchBook, the company, which has raised over $1.5 billion in funding and was last valued at around $3 billion, is now stepping into artificial intelligence; It’s a strategy heralded by new CEO Michael Tannenbaum, who left his role as COO at Brex to join AI. firm. “I thought this was something that could really change the way fintech businesses operate,” he said of his move.

The key AI product he pushes for is to help “look and compare” examples in the lending process. He gave the example of a property-level description, which is a unique description of the asset that must be exactly the same across many legal documents. Traditionally a person would have to look at 60+ pages to make sure the description is the same. Tannenbaum said its new features greatly reduce manual labor and the time needed to verify documents. He said this is an example of AI “eliminating the costs of complex processes.”

He said that given the personal information in loan applications, the company had to go back and forth with OpenAI to make sure its confidentiality agreement was robust and that “the models were not trained in a particular way with our customer data.”

Although this feature runs on GPT-4, chief data officer Ruben Padron emphasized that the company has made it a priority to create model-agnostic systems. “We are constantly testing and evaluating different models as they come out, almost weekly, sometimes daily,” he said. Their system “gives us a lot of flexibility that allows us to quickly and dynamically pivot to the vendor that offers the highest performance.”

Padron sees many more AI offerings in Figure’s future, emphasizing that the more they can automate the loan application process, the less likely there will be errors or bias. “We’re really trying to reduce cost, eliminate manual work, reduce bias,” Padron said. “This is very much a journey. “This is not a destination.”