Optical character recognition (OCR) has been around for a long time, although it hasn’t always been what it is today. OCR — which is technology that can recognize text in a digital document — and the document-automation technologies that use it have significantly advanced over the past decade.
These tools are now giving mortgage lenders access to a wealth of insights that enable more efficient origination processes than ever before. According to recent research from software services company Infosys, OCR can reduce loan processing times by up to 50%. Lenders and the originators who work with them will want to understand how this technology is helping the industry make high-quality lending decisions in less time.
Once a prospective homeowner has applied for a mortgage, the lender conducts an underwriting process to review the application. During this period, the loan originator manually evaluates the applicant’s credit, employment history, income and debt ratios. Then the lender considers other factors such as the home’s fair market value and the applicant’s eligibility for the type of loan being sought. This underwriting process typically takes up to three days.
Underwriting can take much longer, however, if the lender has a staffing crunch or if an unusually high volume of loan applications is streaming in all at once — a challenge that many lenders have experienced over the past year. The applicant’s loan profile also could include some unusual complicating factors, thus increasing the overall turnaround time from days to weeks. Once the lender grants conditional approval for the loan, the process might have to be repeated if the application needs additional supporting documentation, which extends the review process even further. As delays mount, an applicant might be tempted to turn to another lender.
OCR can help a lender determine an applicant’s creditworthiness much more quickly, which in turn enables the lender to quickly scale its operations up or down as needed. For example, the lender can process a higher volume of loan applications and onboard new applicants much faster without having to hire more staff or worry about process bottlenecks that could result in a poor client experience.
A lender can use OCR to gather data more rapidly and effectively than with manual, human-powered and often error-prone processes. Furthermore, a full-service solution can not only use OCR to collect and digitize this information, it also can then analyze and provide insights on this proprietary data. From there, the lender can easily make more timely and accurate decisions about a mortgage application. At a higher level, the lender can even use these technologies to improve its underwriting standards.
A mortgage lender typically draws from various documents, many of which may have come from multiple sources, as part of its qualification process. These documents often arrive in different formats, making it more difficult for the lender to standardize all of the information within them.
According to Infosys, an average of more than 25 types of documents are involved in the retail mortgage origination process, and lenders spend more than 60% of their time on manual operations due to the complexity involved in managing them. When lenders have to work with such large sets of unstructured data, they cannot easily automate any of the processes required to process a mortgage application.
As a result, these companies have difficulty modernizing their loan origination processes or competing with lenders that can deliver faster mortgage approvals. OCR can help lenders overcome this common obstacle by taking raw information from multiple sources and turning it into structured data.
Once OCR has properly formatted the information needed to evaluate and process a loan application, a lender can use an automation solution to quickly generate the insights needed to accelerate their loan origination processes. This way, lenders can rapidly process, underwrite and approve loan applications. They also can scale these processes, using their newfound agility to help more people apply for and receive loans in less time. This is just one way that artificial intelligence is affecting the mortgage industry today.
The loan file review is one of the most important parts of the mortgage process. Because it is manual, it is labor intensive and errors can easily crop up along the way, negatively affecting future steps in the loan process. During the file review, a lender must read through an applicant’s loan file, sort and verify all documents within it, and extract data from it. All of this manual work can quickly translate into a slow turnaround time, diminishing borrower satisfaction and revenue generation.
OCR can provide lenders with timely assistance in this area by uploading documents to be scanned. This enables lenders to digitize data much more easily and bring it together in the proper format. Then the lender can deliver all of the necessary information directly to a lending automation solution.
This eliminates the need for a lot of painstaking, manual labor and allows the approval department to focus on the most important task — carefully reviewing the loan file. With easier access to the right insights, these employees can make faster decisions. The mortgage lender can then accelerate its loan approval process, bringing more revenue in the door and securing a strong competitive position in the market.
As Infosys notes, OCR powered by artificial intelligence and machine learning can address many of the challenges that mortgage lenders are experiencing today. OCR now scans characters in documents with up to 99% accuracy, empowering lenders to automate time-intensive and error-prone documentation processes. By onboarding a full-scale automation solution that utilizes modern OCR techniques, lenders can streamline the onboarding process, standardize loan documents and automate the loan file review process. ●