(This article by Scott Nelson originally appeared in Volume 30 - Issue 3 of the AACFB Quarterly Publication)
Equipment Finance is a business that feels like running on a treadmill set at twelve. And the roughly 12,500 finance brokers operating in the space are trying to help lenders keep up.
The business runs at a three-minute mile pace. It’s driven by volume – sales volume. With the average product life of only three to four years, lenders have to “re-sell” a third of their book of business every year. A billion-dollar lender must sell over $350M in new deals each year just to hold the line, more if they are trying to grow.
That’s a lot of deals and the front-end, intermediary partners generate tremendous amounts of data to help bring those deals to lenders. But most of this data ends up in the origination systems of the lenders who keep and use it to help them make better decisions. Now AI models are helping lenders learn from this data faster by predicting outcomes like:
- What rate and term will be best for the customer of a given deal?
- What type of financing product best fits our risk model for this customer?
- Which lending portfolio best fits this deal?
Lenders can deploy AI products that answer such questions because they typically keep the data needed to model and predict outcomes. Finance brokers working with lenders in digital ecosystems also have the opportunity to take advantage of the technology and gain the same competitive advantage that lenders have. Even more important, make better decisions, faster.
A typical origination workflow can be modelled with five major steps. The first two stages offer great opportunities for increased efficiency and customer satisfaction using AI.
First is “the customer.” In this stage the customer provides information on who they are, what they need, and why they should be able to borrow funds for their purpose. In the next stage, this data is recorded, organized, and submitted by the finance broker to several lenders who may be interested. This is a challenge today because finding the right lender the fastest is key to winning a deal, but many lenders now require information to be entered into digital portals. These portals automate the ingestion of important data into their systems and promise to speed up the match making process. However, many are frustrated with lender portals because there are inconsistent data requirements, they require manual entry, multiple document uploads that also differ from lender to lender, introduce delays of feedback on applications, and have process complexities that require regular visits to remain efficient.
The origination workflow generates a lot of data, so lenders, particularly those who have digital platforms, are more and more likely to require that both the intermediary and customer provide this data in online systems. Digitally mature lenders today are using everything from app-only, direct-to-customer automation, to AI-based prediction machines to make better underwriting decisions and make those decisions faster. But because origination volume is so critical to the business, much of that data is initiated by finance brokers – it often runs through their hands first.
Finance brokers can take the same approach to data as lenders and realize the benefits automation can provide.
To get started, brokers can start by implementing three simple strategies.
1. Capture and keep your data.
To start, finance brokers should begin using their existing systems that capture and keep workflow data. Customer Resource Management (CRM) systems like Salesforce, HubSpot, Dynamics, and Monday.com can all be used to map and digitally record the origination workflow. Don't just give away your data to others via applications, do your best to keep it.
Obvious types of data to keep are those required by lenders such as customer identification and credit data. Digital systems can automate the transmission of this data into the variety of portals used today.
Another recommendation is to capture outcome data, or the results of deals completed. Who were the matched parties? What terms and product types do each lender prefer? What kinds of equipment and from what industries are preferred? How long did closing the deal take? How far was the customer from the lender? Think of every outcome that is important to your success and find a way to measure it.
Again, the main message here is to identify and keep your data. Keep all of it. Keep it for internal use. Keep it for reuse. Keep it to learn and get better.
2. Aggregate, organize, and use your data - every day
Once you have identified the data, then identify the sources. Find the data streams that drive your workflow. Peter Drucker famously said, “You can’t manage what you don’t measure,” so think hard about measuring anything and everything you want to improve.
Now aggregate the data streams and organize it in a data mart with a schema that supports business intelligence. This can sometimes be built organically within a CRM ecosystem, but more often can be best done in a cloud-based middleware platform that integrates to other systems via APIs and ETL functions. The business intelligence system can help manage the sales funnel from initial contact to funding. Each step of the process is time stamped so that productivity is measured by stage, by rep, by equipment class, by vender, and even by lender. Reporting functions allow your operations to see and address bottle necks and less successful business streams – that way you can improve what you measure.
3. Use AI to learn from your data and make better decisions faster.
Once captured and organized, you can learn from the data and use it to accelerate productivity and drive improved performance. The speed and volume of data generated by finance brokers is a natural fit to machine learning (ML) and AI tools. Historical data with outcomes like deal terms and lender partners can build models that can predict which lender is a fit for a given customer’s deal. (Figure 2 shows how automation can simplify the workflow and accelerate it using prediction models.)
A rate or terms predictor can identify which deal the customer is likely to accept before it is sent to a lender. A lender predictor identifies the best lender to fund the deal based on both inputs and term predictions, reducing the time and effort needed to make the match. The deal comes back from the lender in a form the customer anticipated, and funding is assured. Both the speed and accuracy of the business operations are improved.
Finance brokers generate more data faster than any other part of the equipment finance life cycle. As a result, they are in the unique position to take advantage of the decreasing cost and increasing scale of machine learning and AI technology. Finance brokers who engage with their data – can capture it, organize it, and use it in daily operations. Learning from it - will help them step ahead of the competition by making better lending matches faster.