outsource mortgage underwriting services

AI is No Substitute for Underwriting, But It Makes Underwriting Remarkably Efficient

Underwriting Sep 17, 2021

An underwriting job was a daunting task in the past primarily because it consisted of a disjointed process marked by several manual and repetitive tasks. Cut to the present, automation has freed underwriters from being locked into legacy processes of the past. It has allowed them to move away from the conventional linear processes toward exception-based processes. This has introduced a host of benefits some of which include easy and enhanced scalability, delegate junior underwriters with more responsibilities, and free senior-level underwriters to focus on more complex issues. Does that mean AI can overtake underwriting any time soon? We believe no, because underwriting is too complex a process and AI technologies will fall short of meeting the facultative application it demands.

What has Led to the Rise in Complexity of Mortgage Underwriting

New Mortgage Rules Requiring Intense Verification

The new set of mortgage rules imposed in 2014 had a substantial impact on the way lenders originated home loans. For instance, the Ability-to-Repay rule requires lenders to thoroughly verify the financial ability of the borrower to repay the loan. The cumulative impact of many such rules resulted in underwriters investing more time in the process.

A recent survey of mortgage consumers revealed that about 51% of consumers prefer a bank using machine learning capabilities rather than humans to approve loans.

Rising Complexity of Appraisal Standards

After the 2008 housing crisis, mortgage companies were mandated to follow stringent appraisal guidelines. The flexibility that lenders enjoyed before the crisis had given way to more complex standards at a number of levels. This slowed down the underwriting process and made it more time-consuming

The Need for Additional Information and Documents

Procedures such as having a “clear title” to the property are important to getting a mortgage loan closed. Underwriters need to ensure that judgments or liens on the property, are cleared before close approving the loan. To make sure things are on the right side, underwriters often request additional documents, including letters of explanation from borrowers, which invariably prolongs the process.

How Does AI Make the Underwriter’s Life Easy?

By assigning time-consuming document reviews to OCR applications that can find patterns in property images and using natural language processing tools for mining texts for word sequences, underwriters are already realizing tangible gains of AI.

Assistance in Decision-Making

While evaluating the condition and market value of a property, underwriters make a series of easy and complex decisions, some of which require human discretion. While AI is yet to evolve to a state where it can handle all judgment-based decisions, applications with mathematical algorithms are very handy for underwriters in making complex and quick decisions. AI algorithms can be modified to consider and learn from all types of underwriting cases for reduced human involvement.

Unassisted Data Analysis

For automated loan underwriting, AI tools with analytical capabilities can collate huge chunks of information from various documents and instantly create a detailed loan package. These tools can analyze questionnaires, names, vehicle records, and other details of the applicant; locate signatures; classify pages; perform complex calculations; and generate reports. They are designed to eliminate redundant data entry and manual errors, reducing the processing time and errors.

Case Documents Scanning

A typical mortgage loan consists of thousands of pages that a underwriter needs to scan thoroughly, that consume precious hours of their work time. This is where AI-based page classifiers come to the rescue. These applications can scan huge volumes of case files, extract relevant information, and pass it on to the underwriter. AI document scanners help underwriters to spot any wrong document that might be inserted into the bunch of documents, mixing in the information of another applicant.

Stringent Adherence to Underwriting Rules

A modern AI-based tool like MSuite, for instance performs as a robust and dynamic rule engine that underwriters can use to manage thousands of mortgage rules seamlessly. MSuite can easily be customized to the unique conditions of a particular case so that it covers just those rules that apply to the scope of that work. The rule engine of the tool automates distinct underwriting tasks such as thorough file reviews for servicing and underwriting of certain types of loans with a specific borrower profile.

AI Complements the Capabilities of Underwriters, it Doesn’t Replace Them

Over the years, Artificial Intelligence has evolved as a reliable tool in the mortgage arena to carry out operations efficiently and cost-effectively. Today, most lenders are investing heavily to tap its unexplored potential for reducing the dependency on the human underwriters, cutting down on underwriting errors, expediting loan closures, and improving borrower experience. As lenders add complex verification capabilities and alternative data to their workflow, AI will be paramount to meeting the expected standard of fast and efficient credit decision-making.

The closing rate of mortgage loans is calculated on a 90-day cycle instead of a monthly basis since the closing time for most loan applications is usually one-and-a-half to two months. For higher profits, lenders must shorten their loan lifecycle and therefore, the underwriting process.

AI hasn’t and cannot supplant the role of human underwriters, but it can assume the lion’s share of the easy stuff so the underwriters and analysts can focus better on the other areas requiring human judgment. At present, most automated processes involve some degree of human monitoring and checkpoints for validation of information and continuous training of the models so that the higher efficiency can be realized at the earliest.

A large mortgage lender in Massachusetts implemented OCR technology with machine learning capabilities to calculate the income of loan applicants, validate stipulations, and verify loan documentation. Earlier, the underwriting process was carried out manually and took multiple people hours. Now it only takes minutes for the underwriters to do the job with much higher accuracy, and the company doesn’t have to invest in additional headcount as it grows.

As mortgage companies regain their lost ground in the post-Covid era, AI would be their magic wand to orchestrate a comeback. The technology augments the underwriting process for underwriters to capitalize on all data points for optimal outcomes.

In a survey of mortgage lenders, 42% of participants told that their primary objective with AI/ML is to improve operational efficiency. Almost 41% said that they expect such technologies to improve the borrower experience.

Who We Are And Why Our Expertise Matter

Expert Mortgage Assistance (EMA) is a trusted name in mortgage back-office support services. The company leverages its proprietary automation tool to deliver a range of mortgage services. The tool doubles up as a indexing engine, data extraction engine, rule engine and reporting engine to make loan processing job a breeze. The tool has assisted its clients manage 100% or more volumes with ease, has provided 70% quicker return times, and at 50% of cost savings on repetitive manual process. Get in touch with us to know more.   

No Comments

Leave a comment

Your email address will not be published. Required fields are marked *