GARY SNOWBERGER
President Cybertek-Cogensys Dallas
THE ORIGINAL BENEFITS OF artificial intelligence in mortgage underwriting were the increased integrity of decisions, increased communications and knowledge between work groups, and collapsed business transaction cycles. These benefits still hold true. The primary difference today is that these benefits were originally isolated for the underwriting event. With the advent of other emerging technologies wrapped around AI technology, the industry is now positioned to streamline and fully integrate the entire mortgage transaction process -- from data collection through servicing. Marrying and integrating AI systems with the front-end data collection systems, EDI networks, credit bureau interfaces, servicing systems, and document management systems positions the mortgage lender to magnify and significantly enhance the original benefits of AI. The end result is that the lender's competitive advantage grows and customer service is enhanced.
JOHN WOLF
Executive vice president Computer Power Inc. Jacksonville, Fla.
WE BELIEVE that if used properly the new automated underwriting systems will change the velocity with which you make decisions. By having access to quality data at the point of sale, you can make decisions earlier in the process. This would enable you to define which loans you pay more attention to.
Similarly, AI adds value to change management or workflow management to distinguish what areas pose a problem or raise questions. If you look at the traditional underwriting process, we wait until we've done everything to put it to underwriting and then if we do anything wrong or you've got something incomplete, you have to go back into part of the process. If we impose this review element earlier, then we might be able to do things more completely at the inception of the process rather than at the end.
GREGORY A. SAMP
President Sibley Mortgage Co. Rochester, N.Y.
ONE OF THE THINGS that the industry is looking for is something that will help reduce processing time, which in turn would cut costs.
AI offers a way to make the loan review process more even-handed. I think it is particularly important to the industry that organization such as Freddie Mac are coming out with their own programs. In reality, what most of industry is doing is trying to underwrite to secondary market requirements. This will mean having a secondary market agency take a look at the loan, run it through a system and then indicate whether it is acceptable or not.
AI systems can also help us to tackle the issue of fair lending. Although this hasn't been one of primary thoughts on everybody's mind, I think it's a great color-blind way to at least take a first pass at underwriting. I don't think we ever want to be in the situation where the computer says yes or no and there's no way for individual to do anything about the noes. With any system, the loan officer has to be able to look at the application and see if can something can be worked out.
JEFF BUTLER
President Countrywide Credit Industries Pasadena, Calif.
AI will play a major role. It will improve consistency, and the ability to run interim decisions. It will also enable lenders to generate instantaneous decisions based on the data submitted. When you take an application, the customer gives you a lot of information. Granted it's not all verified at that time. If the sale price and appraised value are reasonably close to each other, you could theoretically take the application information and run it through your system fight away to find if there are any weak points.
As the loan progresses and more data [are] validated, you can, at each juncture, deal with issues the system raises and how it relates to the product the person wants to buy. You can make decisions along the way rather than wait for weeks. If a customer isn't going to qualify for a particular loan type, you might be able to switch them to a different product and do it proactively rather than waiting until the end.
The other important factor is that Freddie Mac and Fannie Mac are creating their own AI systems that will say, in effect: "If you underwrite it through my guidelines I'll buy the loan, no questions asked." Basically, there are going to be a lot of automated underwriting systems out there and as an industry, we need to come together and create an electronic data interchange standard that says "These are the basic data elements to underwrite loans."
PHILIP C. FREEMAN
President Data Select Systems Inc. Woodland Hills, Calif.
THE MAIN ROLE that AI can play is to increase productivity and give a "go" or a "no go" decision based on profiles of successful underwriting decisions. I also think it can play a large role in monitoring policy compliance for each lender. You can put all your policies out there and it can go through and check those policies and really make sure you are in compliance; and if you aren't, it can push you back in.
AI in automated underwriting systems can also be used as a great senior leaders The decision trees of more senior lenders can guide more junior leaders.
ROBERT WARRINGTON
President Old Kent Mortgage Co. Grand Rapids, Mich.
ARTIFICIAL INTELLIGENCE in traderwriting will become a way of life here shortly. First, it will greatly reduce the time it takes to approve a transaction. Second, it will offer significant cost benefits to the borrower and to a fairly significant extent, it should reduce the cost of loan origination.
I believe that a high percentage of loans -- perhaps 50% to 70% -- could be approved subject to property appraisal at the time the application is taken. This means adding on an electronic data interface between credit bureaus, title companies, and, ultimately, tax assessors offices to determine the value of the properties. We will wind up approving the whole transaction right there upon application.
