Lending fintech Ocrolus is slashing time spent on mortgage underwriting and income review with AI, the company said. 

Ocrolus uses AI to extract data from documents, identify mismatches across documents and analyze borrowers’ financials, reducing mortgage underwriting times by as much as 30%, General Manager of Mortgage Nadia Aziz told FinAi News.  

“In the mortgage world, there are more than 2,000 different types of documents that you may come across,” she said. “Instead of spending hours and days on certain things … you’re reducing the cycle time.” 

Ocrolus’ AI solutions also are cutting income review times by as much as 90% for some clients, including American Federal Mortgage, according to a June 15 case study.  

Prior to deploying Ocrolus’ platform, American Federal employees had to manually review 12 months to 24 months of bank statements, Gail Stibitz, senior vice president of underwriting at American Federal, stated in the report.  

The mortgage lender now uses AI to verify income and has cut the time spent doing so to minutes instead of hours, Stibitz said.  

Ocrolus also recently rolled out an AI-driven “conditions” tool that allows underwriters to quickly surface Fannie Mae and Freddie Mac guidelines for certain loan products, Aziz said. 

This helps underwriters obtain the required data and documents “to originate the loan that Fannie or Freddie will be willing to buy,” she said. 

For example, some loan products might necessitate two years of tax returns or have third-party processor stipulations, Aziz said. 

“An underwriter spends an inordinate amount of time going through each document and saying, ‘what is here, what is not here,’” she said. 

Improving borrower experience  

Not only does AI enhance underwriting efficiency and cut costs, it also improves the borrower experience, especially for first-time home buyers, Aziz said.  

In many cases, simple discrepancies, such as omitting “Inc.” from an employer’s name in some documents while including it in others, bogs down the process and frustrates borrowers, she said.  

AI quickly flags these inconsistencies and eliminates back-and-forth conversations over seemingly minuscule requirements, she said, noting that the gaps between these conversations often take days or weeks. 

“If we’re able to get the documents, extract the data, do the income calculation right away up front … we can let them know what more we need right away,” Aziz said. 

Register here for the FinAi Lending Summit, set for Oct. 7-8 in Las Vegas.