Online lender LendingClub is deploying AI throughout its operations, including to improve underwriting. 

“At this point, there’s probably not a department in the company that is untouched [by AI] in some way,” LendingClub Chief Executive Scott Sanborn said during the company’s fourth-quarter earnings call on Jan. 28. “I think we have over 60 initiatives underway across the company.”

In addition to underwriting, the lender has deployed AI for: 

• Fraud detection;  

• Collections;  

• Document verification;  

• Marketing; and  

• Customer service. 

AI underwriting 

Further, LendingClub’s customized machine learning underwriting model “delivered 40% to 50% better credit performance versus our competitive set,” Sanborn said. 

The company’s AI-driven underwriting model can predict a consumer’s potential default rate 12 times better than a FICO score, according to its Nov. 5 investor day presentation.  

In the presentation, for loans originated by the company, it reported delinquencies of: 

• 2.2% for a FICO score of 660-719, compared with 4.2% for its competitors; 

• 1.4% for a FICO score of 720-779, compared with 2.2% for its competitors; and 

• 1% for a FICO score of 780-850, compared with 1.6% for its competitors. 

During Q4, LendingClub reported: 

• Loan originations of $2.6 billion, up 40% year over year; 

• Total revenue of $266.5 million, up 22.7% YoY; and 

• Total assets of $11.6 billion, up 8.8% YoY. 

Register here for the inaugural FinAi Banking Summit, taking place March 2-3 in Denver. View the full event agenda here.