Nearly every day, there seems to be a new bank adopting artificial intelligence. Yet, if you were to ask 20 different banking professionals on what AI is, you’d get just as many opinions.
The industry’s lack of consensus around a definition stems from the fact that AI is constantly evolving. In fact, artificial intelligence goes back as early as 1956 at the Dartmouth conference.
Since then, AI has become ubiquitous from defeating grand-master chess champions to helping you order from your favorite restaurant. As a result, it’s understandable that due to its wide variety of applications, AI may seem challenging to define.
According to Webster’s dictionary, it is the capability of a machine to imitate intelligent human behavior. So, what is artificial intelligence within the context of the mortgage industry?
Many mortgage professionals tell us they have AI-enhanced technology but are disappointed when it falls short in achieving the expected results.
AI must meet one of two conditions for a solution to be considered AI-enhanced. The machine must perform a task as well or better than it would if done by a human. The alternative is that it must hold the potential to attain or surpass human precision.
This means that if a human is needed to fix a significant percent of the machine’s mistakes or that the machine isn’t significantly cutting the time it takes to process a loan document, the technology isn’t truly AI-enabled.
True artificial intelligence must improve existing business processes and exhibit intelligent behavior. When we meet with companies in the industry, many will think they have an AI solution. Yet, what they implemented was a solution that creates more inefficiencies rather than it solves.
How should you evaluate solutions in mortgage? There are several criteria that can be tested during a proof of concept.
- Speed – Does your solution process more loans at a faster rate? A true AI solution can reduce processing time by at least 80 percent.
- Accuracy – Does your solution dramatically reduce the error rate such that you can be certain the machine classified and extracted the right information? Without some type of validation method in-place, you might find that correcting the machine’s error tacks on un-needed time to the loan process.
- Business process – Does the solution improve your business process or make it more frustrating? An AI solution should allow your company to move employees from tedious tasks to more strategic ones, that will ultimately make your business more profitable. If your employees cannot automate much of what they’re doing, then it’s not a true AI solution.
Learn about our AI-enhanced loan processing solution, Trapeze for Mortgage Automation.