MReport February 2019

TheMReport — News and strategies for the evolving mortgage marketplace.

Issue link:

Contents of this Issue


Page 31 of 67

30 | TH E M R EP O RT FEATURE T here are significant changes taking place to mortgage lending rules. To proactively stay informed of the changes, be in compliance, and remain competi- tive it will be important for both bank and nonbank lenders to embrace leading-edge technology solutions that are based on artificial intelligence (AI) technology for processing loans and maintaining compliance. Using both AI and machine learning (ML) can help lenders process more mortgages in less time while conforming with all of the federal, state, and other rules. Mortgage Rules Are on the Rise A fter the 2008 housing crisis, government-sponsored enti- ties (GSEs) such as Freddie Mac and Fannie Mae enacted a set of standard rules that must be followed when processing a loan. There are also state and local rules with which lenders need to comply, and government-backed and Veteran Administration (VA) loans also have specific rules. Additionally, the Home Mortgage Disclosure Act (HMDA), originally enacted by Congress in 1975, was updated with changes that were signed into law by the president on May 24, 2018. For HMDA, 47 different data points need to be monitored for key documentation in a loan package. It is essential to identify the new critical data points re- quired and take action to provide the additional HMDA data, or you could face fines by the CFPB. Keeping up to date and compliant with HMDA should be a priority for lenders in 2019 and beyond. In a tight mortgage market, there are naturally fewer mortgage loan applications. To stay competi- tive and profitable, the industry is responding to this lack of loans by offering different, nontraditional products to try to generate revenue. Examples include home equity lines of credit (HELOCs) and non-qual- ified mortgages (non-QM). These products have their own unique set of rules, which again increases the overall number of rules that a lender needs to be aware of and meet for compliance purposes. The bottom line is that there are more rules today, which adds complexity to loan processing and makes compliance more challeng- ing. Accelerating Loan Processing (and Rules Processing) with AI H ow can lenders stay on top of all these rules? Using AI and ML can help automate a significant amount of mortgage loan processing, which includes rules processing and compliance. This not only creates a faster system, but a more accurate one that can help reduce errors and improve compliance. AI allows much of the work that was pre- viously done manually to now be automated, including reading through hundreds of loan docu- ments and verifying informa- tion against the particular set of rules for that loan. With ML, rules can be pre- loaded into the system and "learned" in advance. When new loan documents are entered into the system to begin processing a mortgage, it will use vision-based learning technology that scans the documents, collects the right information, and then matches it against the correct set of rules for that loan for verification. The system will know to check and verify for only the specific set of rules that apply to the loan that is being processed. As the AI-based system identi- fies missing information and errors, it can organize and classify documents by error level, identi- fying and isolating all of the risks. And, lenders can classify docu- ments as "accepted" or "denied," and then organize the denied documents based on why they have been denied. This allows lenders to prioritize their efforts for correcting errors and gather- ing additional information. All of this helps lenders in their efforts to comply with federal, state, HMDA, and other rules. The Benefits of AI— Speed, Accuracy, Customer Service O ne key benefit to imple- menting AI is how fast all of this processing, verification, and classification can be done. There are more than 600 rules that could apply to a loan, so from a volume and capacity standpoint, there is a significant advantage to automating loan and rules processing. In addition to speed, accuracy is also im- portant. Having a human read through hundreds of documents can lead to fatigue, which opens up the potential for errors. These errors could lead to com- pliance issues—and fines—so it is important to be as accurate as possible. AI technology can run on a near constant basis without fatigue, so there are advantages to deploying this technology when processing loans and the rules associated with them. Improved speed and accuracy can lead to other benefits too. First, productivity increases—lend- ers can process more loans in less Welcome to the Machines AI and machine learning could be transformative for mortgage —if they're leveraged properly. By Jill Jones

Articles in this issue

Archives of this issue

view archives of TheMReport - MReport February 2019