APPROVE has been called a "disruptor" in the equipment finance world in part because we are not a traditional finance company. We are a tech company working to solve problems for equipment sellers and buyers. But what is the right balance between tech and human capabilities in this work? And how can machine learning be used to improve the equipment sales journey? Our Chief Technology Officer, Wellington Souza, talks below about how APPROVE approaches this exciting work.
Q: How does APPROVE use machine learning to speed up equipment financing?
Wellington: We are a tech company, but we place a high value on the human-to-human interactions and personalized customer service that deliver a superior equipment financing experience. So, we wanted to leverage machine learning in a way that would speed up the manual portion of the financing process, and also unburden our financial operations team members from cumbersome tasks that took time away from their interactions with customers and could be automated.
The solution we built uses machine learning to match financing applications with their best-fit lenders—a step in the financing process that is not a human touchpoint in any way. Machines are not going to replace the highly trained personnel we have working with our customers—they are going to help them be better and more efficient at their jobs. And humans will always be necessary to share and analyze the extra anecdotal information that often accompanies applications, especially for larger transactions.
Q: What makes machine learning a good tool for matching finance applications with the best lenders?
Wellington: A major advantage of APPROVE is our lender network, which is made up of lenders that specialize in a wide range of industries, business types, equipment types, loan types and credit profiles. We can accommodate all types of applications, and the machine learning tool we have built allows us to instantly identify those lenders in our network most likely to approve a particular application—and to offer the most competitive finance rates and terms.
APPROVE has such a vast and dynamic lender network that it would be impossible for people to recall, process and analyze so much lender data and equally impossible to notice and keep up with internal and external changes affecting each lender’s underwriting policies and tendencies. Machine learning can instantly detect trends at the individual lender level that a human analyst would not easily or quickly recognize. The system observes and tracks these changes in real-time, and uses them to constantly update and improve the algorithm.
Q: What factors is APPROVE’s algorithm optimizing for when it seeks the “best-fit” lenders?
Wellington: There are two main factors:
If an application is really strong and likely to be approved by every lender in our network, machine learning will predict which three lenders are most likely to offer the lowest financing cost and most favorable terms. If a buyer is looking for a particular type of loan, the algorithm will eliminate lenders that don’t match the desired criteria.
Q: How does all of this lead to faster approvals?
Wellington: Once our machine-learning algorithm selects the best-fit lenders, our technology instantly sends the applications to those lenders. This automation is the key to lightning-fast financing. Before we used this technology, it could sometimes take an hour or two for our operations team to review applications and publish them to our lenders—longer if the application came in near the end of business hours. Those days are over. In conjunction with some new automated underwriting and instant approval technology we recently launched, our machine learning technology is capable of submitting applications to optimized lenders and receiving approvals in a matter of seconds. But, even if an application doesn’t qualify for an instant, automated approval, the applications are being optimized and distributed to best-fit lenders almost instantly. It no longer matters when an application is submitted by a customer, we’re going to get it in front of the right lenders immediately. Apps that come in close to 5 p.m. (end of the workday) and are not approved through our automated underwriting process, now get approvals before 10:00 the following morning, showing that the automation we have built ensures that deals do not lose momentum.
Q: How does this speed boost equipment sales?
Wellington: Our data clearly show that early engagement and speed of delivering finance application approvals are the two most important factors in increasing conversions. Early forward momentum drives conversion rates up significantly. This is a huge benefit to the equipment sellers on our platform. The more efficient our algorithm gets, the more deals they are going to close. Lenders also see their business grow, as the speed with which we deliver applications to them also drives their conversions up. And because our machine learning has already selected those applications as a good fit, our lending partners spend far less time working applications that would eventually get declined. Buyers experience a much smoother, quicker sales process, maximize their chances of approval, and receive the best possible deal terms.
Q: How often are you upgrading this tool?
Wellington: Ultimately, we are here to make life better for our customers, who are the lenders, buyers and suppliers using our platform. We are constantly asking them what problems are they experiencing and what can we do better.. We listen to their feedback and we respond quickly. We’re also very data driven, and our data can expose bottlenecks or productivity drops before they reach the point of a customer complaint. We are constantly developing solutions and implementing improvements to remove roadblocks to our customers’ success and satisfaction.
Q: What motivates your team to do this work?
Wellington: Our development team is top-notch, and we really don’t think there’s anything we can’t build. The spirit of innovation comes from the very top at APPROVE. Our CEO, Robert Preville, is an engineer and a developer who has a dynamic and unique approach to problem-solving. He is definitely the primary influencer of the innovation mindset. His vision is inspiring to all of us—developers, sales executives and our financial operations team. We have an office culture that emphasizes fun, team-building and taking joy in being able to live and work in a place as cool as Wilmington, N.C., and that energy makes us great collaborators. When we bring all of our backgrounds and skills together around a problem, it always leads to creative ideas. We just keep building things to bring those ideas to life for the benefit of our customers.