Transforming Decision Making in Middle Market Lending

From origination to funding all in a millisecond.

According to Andrew Yang, American entrepreneur, the founder of Venture for America and 2020 Democratic Presidential candidate, we are in the early days of the “greatest economic and technological transformation in the history of the world.” Automation is disrupting industries across the globe. From transportation to manufacturing, healthcare to finance: No industry is immune from radical transformation. Yet several pockets within these industries have been resistant to disruption. Market participants in these segments know that it is just a matter of time, and the clock is ticking.

As a finance company, we at CapX Partners know that one such segment is middle market commercial lending. Unlike consumer and small business lending, where credit decisions are made in milliseconds by newcomers like Kabbage and incumbents like CapitalOne, middle market commercial lending remains slow and inefficient, taking months to close a single $10 million transaction.

An era of disruption

The economic impact of middle market companies cannot be underestimated. There are roughly 200,000 companies in the U.S. with revenues ranging from $10 million to $1 billion, most of which are closely held or family controlled. Such businesses produce roughly one-third of the U.S. GDP.

Financial institutions already lend trillions of dollars per year to middle market businesses, and there is huge growth potential for those that can quickly and accurately originate, underwrite and fund transactions. Entrepreneurs and innovators are doing what they always do when there is big money on the line: They’re digging in, looking for the bottlenecks, finding, productizing and monetizing solutions. Hello, disruption!

It’s already happened elsewhere

Many financial institutions are willing to lend $5,000 -$250,000 based on data from Experian, PayNet, FICO, and others. More sophisticated credit evaluation solutions like Ocrolus include digitized bank and tax documents that provide deeper insights into the financial condition of the borrower. Digital lending platforms have gone further and added hundreds, even thousands of additional data points to create a more accurate picture of the borrower. These can include everything from an owner’s social media presence to corporate and personal purchasing patterns. These financial and behavior-based analytics generate risk assessment results, and a yes or no response is auto-generated. The speed and accuracy of these solutions create a better experience for both the lender and the borrower. But what happens when a company is looking for $5 million, or even $50 million?

Up next: Middle market lending

In the case of a larger, middle market company, lenders need to rely on financial statements to gain a more confident assessment of the borrower’s ability to repay its debts. Therein lies the challenge. First, no two companies produce financial statements that look exactly alike. So a manual process is required to move the financial information from analog financial statements into digital systems that can calculate ratios, identify insights and facilitate credit decisions. This manual, bespoke process is slow and error prone.

Second, balance sheets and income statements, even with the typical copious footnotes, provide a wealth of data, but don’t provide rationales or explanations of the data. We might see $100 million in revenue at the top of the income statement. But how is that revenue being generated? Is it relatively evenly distributed among a large number of customers? Or is it so top heavy that if one customer leaves, the company structure starts to crumble. These qualitative considerations are not always reflected in financial statements but are important in creating the complete portrait of an organization.

First, automation, then true AI

At present, financial analytics companies such as Fincura claim to have cracked the code, by unlocking the data in financial statements. Combining this digitized financial statement data with the behavior-based information initially designed and used for consumer and small-business lending, auto-decisioned credit processes that target middle-market companies can now be designed with speed, accuracy and scale never before imagined. But automation is not AI.

Artificial Intelligence is based on the premise that machines can learn and adapt from experience, rather than rely exclusively on pre-programmed logic. AI can detect tendencies and predict what a person, or a business, will need or want in the future. Machine learning excels at pattern recognition, with image recognition and speech recognition some of the most common uses. Both AI and machine learning are evolving at a rapidly accelerating pace, and over the past decade, we’ve experienced their impact in numerous ways. One area particularly important to the financial services industry is safety and security. Algorithms can now block fraudulent transactions with a high degree of accuracy by combining pattern analysis and predictive analytics. AI has also improved the customer experience and increased personalization; for example, answering customer questions in real time via a chat bot. And now, AI is poised to improve middle market credit decisioning and make it faster, more accessible and more transparent.

True AI requires deep learning, and deep learning requires millions of data points. It will take time for true AI to completely replace the human involvement in middle market lending decisions, but the path ahead is clear--first comes auto-decisioning followed by deep learning, leading to true AI.

CapX Partners and its parent company, Accord Financial, are working on creating a world where middle market credit decisions are made in milliseconds, using AI and machine learning to combine quantitative and qualitative data in a systemic way to provide the complete picture of a borrower. Our mission, no matter how we get there, is always the same, AI or not: To simplify access to capital so our clients can thrive. Middle market lending is not a winner take all proposition: There will be winners, and there will be losers. We intend to be among the winners.