All companies make poor decisions from time to time. Often, a company’s refusal to embrace a disruptive technology in favor of tried-and-true traditional methods, doesn’t turn out well. Think of Kodak, the once-great imaging empire, inventing the first digital camera in 1975 but deciding not to sell it right away for fear that it would cannibalize film sales. The company filed for bankruptcy in 2012 and now operates as a black-and-white negative of its former glory. Similarly, for a financial institution (FI), a failure to implement new technologies for lending to small and medium enterprises (SMEs) can mean a dim future.
One of the struggles when lending to an SME is the cost that goes into making a lending decision. It is estimated that at regional and community banks, $4-5K in operational costs go into processing each loan under $100K, leaving very little margin for bad loan decisions. In fact, these small loans take nearly as much time and manpower to process as much bigger loans. This is where automation and artificial intelligence (AI) can come into play.
Intelligent automation can be integrated with legacy systems to create robotic workflows from the customer to back office processing. These robots take the routine, repetitive processes, such as performing credit checks and consolidating data across multiple accounts, and make them both more efficient and effective. What was once a very tedious process of new customer onboarding communication between a loan applicant and a bank employee can be replaced by highly personalized interactions based on individual activity post opening.
These robots can speed traditional loan documentation and collection from weeks to a matter of days, all without having to replace or change the existing systems. The customer onboarding and KYC processes can be greatly enhanced from an experience, quality, and cost standpoint. This intelligent automation can drastically reduce a bank’s operational costs of processing an SME loan.
Individually, SME loans tend to be riskier given their higher variance in performance. The lack of standardized financial reporting required with an SME loan compared to larger accounts is one of the reasons predicting loan performance for SMEs is more difficult. How do you predict the performance of similar businesses where, say, one is set up as an LLC and the other as a C Corp? This is another area where robotics and AI can be leveraged. In addition to increased speed and accuracy, the robots can be programmed to use data to perform predictive analytics.
This is where things get really interesting from a lending perspective. Unlike traditional models of underwriting, which focus on only a handful of credit attributes, machine learning can analyze thousands of data points from various sources, which allows for a bank to model credit risk for SMEs more accurately than ever before. These machine learning techniques are able to radically outperform traditional scorecards in SME lending. In the not-distant future, a bank could use robots and predictive AI to 100% automate lending decisions in cases where the SME is under a certain amount and the predictive analytics give the applicant a certain baseline score.
Digitizing a bank’s processes through intelligent automation, combined with machine learning and artificial intelligence, will soon no longer be optional for banks to stay competitive in the future. Online lending platforms touting near-instant loan approval continue to encroach on the role of traditional FIs in the SME loan market. A bank refusing to pivot toward the use of AI in making better, faster lending decisions may not be on par with shelving the digital camera, but it’s pretty close.