The recent CARES Act is challenging financial institutions in a new way. Read my previous article on this topic. The demand on the system, the shifting requirements and liability, the risk of fraud, and thin returns are taking a toll on processing and validation infrastructure. Systems that process commercial loans and small business banking transactions are not well equipped to respond or soothe consumer fears. There are many aspects of the banking and consumer experience that should be improved, but one reigns supreme – automation. This has not been a priority, but it should be.
The landscape for commercial loans is changing. Technology has raised the bar and consumers expect an automated digital discourse. So much has moved online yet the commercial and small business loan origination processes still resemble 1960’s corner store banking. It is riddled with cumbersome paper documents, excel spreadsheets, and approval flows.
If you are a small business lender and you are running manual, paper-based underwriting practices it is likely you are not getting the consistency, auditability, or accuracy you need out of your system. On top of those pains, your originations are probably very, very time-consuming.
Commerical Loans: The Problem
- Striking the balance between Judgmental Decisioning & Automation: There has always been a natural tension between the thoroughness of human-centric underwriting and the efficiency goals of the financial system. Unlike consumer loans, commercial loans do not have the same risk profile, exposure, or levels of protection. The loans are not unregulated, but the spaghetti of regulations does result in lenders baring more of the shared risk. Additionally, it increases risk for error and noise. That risk has been a driving factor in the complexity of the commercial loan process.
- Facing the Incumbent’s Dilemma: Commercial banking is facing an incumbent’s dilemma with no incentive to change. The “That’s the way it has always been done and it’s too hard to change it” rhetoric locks many lenders into a perilous pattern of maintenance. Many have attempted the change only to realize their transformation was just a lift and shift of what they already had. They re-designed the same tired process only now it’s on the internet.
- The Stockholm Syndrome for Bankers: Many bank leaders have designed their loan processes around artifact validation but have not taken steps to truly transform their collection and automation of data. Leaders are captive to their control processes and continue to design systems that value the presence of artifacts over the capture of data. Automating data flows with improved capture, advantaging machine learning (ML) and artificial intelligence (AI) can break this dynamic and free the underwriting process from a check box mentality. Today, the system feeds itself, and in many ways is incented to reject any disruption or re-definition of how it captures, assembles or federates data.
Simple automations can streamline systems and improve data flow at just about any step of the loan process. Below are the four major parts of the loan origination process and some observations around automation:
Step 1: Frankly, get better at collecting information.
In many systems for processing commercial loans, applicant information is collected on a paper form and manually re-keyed to a back-end system. Even in cases where front end portals exist, CRM data is often not pre-populated and workflow engines require multiple manual additions. That experience can be frustrating for both customers and lenders. The manual nature of this process often results in human error, leading to multiple, frustrating, touchpoints to correct basic information.
Start by building an intuitive on-line experience that is adaptive and responsive. One that pulls in data from CRM, pre-populates from commercial and consumer APIs, and eliminates any semblance of re-keying. We have seen the explosion of consumer online originations and improved consumer experiences. We know how to do it. It should be “apple” easy. Once an application is received, a simple rules-based workflow engine should begin the work of classifying applications that are ready for decision and those that require more documentation.
Step 2: Stop spreading in Excel, unleash machine learning, and artificial intelligence.
Perhaps one of the biggest opportunities for automation is in risk analysis and spreading. The aggregating of information that is material to the risk-based lending decision and performing a thorough analysis against the documents provided by the customer, has multiple automation-ready steps.
Build yourself the tooling that automates the capture organization, validation and preparation of documents so they can be scored and underwritten according to governing rules and procedures. Employ a healthy dose of artificial intelligence (AI) technologies such as digital image processing, natural language processing (NLP) and machine and deep learning algorithms. Make the process of spreading go from hours to instantaneous. Moreover, this employment of technology opens up the possibility of pre-scoring and in-principle credit offers.
Step 3: Bring it all together with data in credit presentation & decisioning.
Once the risk is understood and financial spreading has provided sufficient information, the next step is to bring the borrower through the application for credit. For many lenders, this is another manual exercise in collecting several separate, yet related, paper and digital artifacts and organizing them according to the credit lending rules and requirements. This process is often a highly prescribed procedure, followed by rinse and repeat.
Take a customer-first approach and leverage the data you have in your CRM and spreading processes to make the experience as seamless as possible. Remember “Apple easy?” The presentation and decisioning steps should be absolutely frictionless.
Build or buy yourself an automated credit application system that can combine elements across data repositories and flow decisions down according to your loan app procedures. Be careful: There are many applications in the market that facilitate automated credit decisioning, but most tend to follow rigid process-centric design often forgoing modern data management practices and falling short on critical user experience tests. Building custom solutions is typically safer.
Step 4: Decide what should be automated and what should flow down.
In the consumer credit environment, automatic decision making is already commonplace but in commercial loan environments that reality is a little harder to achieve because no two loans are ever the same, so it remains largely human system.
A lack of mirroring aside, there are patterns that can and should be automated especially with high volume / low loan value decisions or for pre-screening and risk assessment process steps. Identify patterns and take action.
Automate commercial loans and reduce risk.
The goal of any lender is the management of risk. One of the foremost reasons for driving automation beyond speed, efficiency and customer experience lies in improved data handling, lineage and governance. Improving data integrity ultimately improves the quality of decisions and reduces risk.
The world is changing rapidly, and it is time the small business and commercial loan processes up their game. Stop lifting and shifting. Leverage modern technologies and re-define your lending process. Not only will you drive cost savings, but these steps will allow you to take control of your data and drive better, more transparent, efficient, and higher quality lending decisions.
We’ve partnered with commercial lenders, both big and small, to automate their businesses and leverage technology for reduced risk. How can we help you?