SBA’s automated loan reviews save thousands of work hours
- By Stephanie Kanowitz
- Feb 05, 2021
By automating low-value, repetitive functions, the Small Business Administration’s Loan Review Tool (LRT) is reducing the four to five hours loan reviewers typically spend on evaluations by one-third, freeing them to focus on high-value work.
Put in production in February 2020, LRT provides a Turbo Tax-like question-and-answer interface that guides analysts through the relevant process associated with the specified review type and the current agency’s standard operating procedures, an SBA spokesperson told GCN. SOPs are the policies and processes that guide the business function, but they can change frequently, making it tough to review loans consistently and efficiently.
LRT can perform automated validations, managing a range of documents and generating government-certified reports, the spokesperson said. Additionally, because the loan review process requires the ability to quickly adapt to new and evolving regulations, rules, forms and procedures, the tool can guide less experienced reviewers through the process.
“Further, the LRT supports SBA’s cloud-first mandate and reaches the agency’s geographically dispersed workforce,” the spokesperson added. The agency has thousands of workers in more than 100 locations nationwide.
LRT runs on SBA’s cloud, which complies with the Federal Risk and Authorization Management Program, and is built on Microsoft’s Power Platform technology, a FedRAMP-certified, enterprise, low-code/no-code solution development platform. It allows the agency to rapidly create and customize applications.
What’s more, LRT seamlessly integrates with other Microsoft services that SBA uses, such as Azure, Office 365, SharePoint and Teams. To that end, LRT uses Power Platform’s integration with Azure and Active Directory to integrate directly into SBA’s security infrastructure; with Office 365 for template-driven document creation and Outlook for internal and lender communication, correspondence and collaboration; and with SharePoint for document management.
LRT is integrated with SBA’s flagship 7(a) loan guarantee program’s SOPs, which change often, making loan reviews loans difficult, the spokesperson said. In fiscal 2020, SBA guaranteed over 40,000 small-business loans totaling more than $22.5 million as part of 7(a).
Loan reviews require expert knowledge of SBA regulations and 7(a) loan documents. Previously, the loan review team consisted of expensive third-party contractors with specialized experience and understanding of the agency’s loan programs. To reduce those costs, SBA wanted to bring the reviews in-house, ensure quality and consistency, increase efficiency and make review results and trends more apparent to managers, according to SBA.
“The system provides automated validations that identify issues that can then be examined for resolution or action,” the spokesperson said. “LRT’s native integration with Microsoft Office 365 environment promotes communication and collaboration with lenders through automated reports, pre-filled Word documents and emails directly from the system. Analytics and data visualization is provided by Power BI dashboards, embedded into the system.”
Because SBA has an annual goal of completing 2,000 loan reviews, LRT stands to save thousands of work hours, according to the spokesperson. Other benefits include standardization of the review function, increased flexibility as SOPs and other requirements change and program insights delivered through analytics and visualizations that are built into customizable dashboards.
The agency used an accelerated agile process with one- to two-week sprints to develop the tool, which went from prototype to production in eight months. SBA worked with FI Consulting, the Office of Performance and Systems Management and its Office of Credit Risk Management on LRT, which was a finalist for the American Council for Technology and Industry Advisory Council’s Igniting Innovation 2020 award.
Stephanie Kanowitz is a freelance writer based in northern Virginia.