What Is Automated Underwriting?

Automated Underwriting Explained in Less Than 5 Minutes

Smiling couple receives the key to their new home
•••

Peathegee Inc / Getty Images

Automated underwriting uses technology to evaluate risk and underwrite loans, and it can greatly expedite the loan approval process compared to manual underwriting. There are both pros and cons to using this type of underwriting.

We’ll clearly define automated underwriting and explain how it works. Then we’ll examine the advantages and disadvantages of using automated underwriting in the mortgage approval process.

Definition of Automated Underwriting

Automated underwriting uses technology known as automated underwriting systems (AUS) to evaluate risk and underwrite loans. It can make the loan approval process faster and easier for both lenders and borrowers—it’s not an understatement to say that automated underwriting brings the mortgage process into the twenty-first century. 

For example, U.S. Bank has explained that mortgage underwriting “can take anywhere from a few days to several weeks.” However, the technology employed in automated underwriting can greatly reduce the time required to evaluate potential borrowers, keeping it to the shorter end of this time frame. How is it able to process a mortgage application so much quicker than the manual underwriting process?

How Automated Underwriting Works

“Automated underwriting is essentially an algorithm making the loan decision—or at least a recommendation for a loan decision,” explained Aaron Dorn, chairman, president, and CEO of Studio Bank in Nashville, Tennessee, in an email to The Balance. “The software or application is fed the key data points, and it then uses formulas to decide whether or not the loan is in the best interest of the bank and borrower.”

Artificial intelligence (AI) is already taking the lead in other financial sectors, like the investment industry. For example, IBM reports that AI-driven trading platforms (without any human intervention) can make anywhere from thousands to millions of trades every day.

But at least right now, humans and AI are working together in the mortgage approval process. 

“AI assesses the eligibility and risk factors associated with income, assets, debt-to-income ratios, loan-to-value, and credit history, and helps determine the documentation that will be required to verify the data,” Melinda Wilner, chief operating officer for United Wholesale Mortgage, told The Balance by email. 

Then, she explained, the underwriter uses the AUS’s findings to guide the final underwriting of the loan. “The underwriter ensures the data entered into the system is accurate, that all the required documentation has been collected and reviewed, and [that] the borrower is eligible for the mortgage loan,” Wilner said.  

Types of Automated Underwriting Systems

Technological advancements have led numerous mortgage lenders to begin using AUSs. For example, in 2015, Quicken Loans launched Rocket Mortgage, the first completely digital mortgage process, and by late 2020, 98% of all of the company’s loans used the technology.

In October 2020, the Federal Housing Administration (FHA) announced its own AUS for FHA loans.

“Fannie Mae Desktop Underwriter (DU) and Freddie Mac’s Loan Product Advisor (LPA) are the two main automated underwriting systems,” Wilner said. “Lenders use either of these systems, or both, to determine if a loan meets eligibility requirements from Fannie Mae or Freddie Mac.”

Does it matter whether your loan is reviewed using DU or LPA? “There are no significant differences between each of these systems,” Wilner said. “It really boils down to which [one] a lender would like to use.”

Automated Underwriting vs. Manual Underwriting

There are obvious advantages to automated underwriting versus manual underwriting

“Automated underwriting systems help to create speed and efficiency in the underwriting process, allowing for better decision-making by underwriters,” Wilner said. “The automated systems can quickly analyze data and do it very consistently, providing thorough feedback and action items for underwriters to complete the underwriting process.” 

Automation can help reduce the paperwork and time involved in reviewing each application, helping lenders to approve more mortgages and use their time more efficiently. 

However, speed isn’t always a good thing, noted Dorn, and automated underwriting isn’t a perfect solution. “It’s designed to benefit the lenders,” he explained, “But that doesn’t mean it’s always ideal for the borrowers.”

Like any other computer program, an AUS is only as good as the data it’s given and the calculations it’s programmed to do.

“So if there is bad data, such as errors in a credit report or missing income, it will automatically produce a bad result,” Dorn said. That’s why he prefers manual underwriting. “Manual underwriting can be slower, but it comes with the benefit of having a human on the other side of the paperwork.” 

He said a smart underwriter can spot and address issues before making a final decision.

An AUS also has limitations that manual underwriters don’t. “For example, manual underwriters can also consider things that may not make it into an algorithm’s calculation, such as community resources and atypical circumstances like natural disasters, hospitalization, a global pandemic, and more,” Dorn said.

Certain scenarios can also prevent the use of automatic underwriting.

“Manual underwriting may be required for certain loans where there are things present in the borrower’s profile that [an] AUS cannot assess,” Wilner said. When this happens, the underwriter will use manual underwriting guidelines—which may differ from AUS guidelines and require the borrower to submit additional documentation. 

Key Takeaways

  • Automated underwriting can expedite the loan approval process, so borrowers learn much sooner whether their application has been approved.
  • Automated underwriting can require less documentation from prospective borrowers than manual underwriting.
  • Automated underwriting is only as good as the information it’s given and the algorithm it’s programmed to use.
  • A human underwriter is required to input the information, review the AUS’s findings for accuracy, and actually issue the loan.