Enterprise Finance and AI: Bridging the Financing Gap and Reaching the Credit Invisibles

By Arvind Nimbalker, Global Head of Product at Tribal

At first glance, the criteria that lenders typically use to determine which enterprises are creditworthy and will receive access to capital seem eminently reasonable. These traditional creditors look at a company’s current balance sheet, their income, and cash flow statements. They may consider the financial histories of the principal individuals involved. Ultimately, they use this information to decide who gets funded and who doesn’t. It’s a methodology that has worked for many years and in many circumstances, but it also leaves a wide berth of companies for whom these criteria are inadequate, termed by some the “credit invisibles."

In particular, companies at earlier stages may not have the historical finances required by the traditional credit calculus. Other companies that have operated for years in emerging economies around the world with less mature financing ecosystems may still not have the kind of traditional financial records required by many lenders. In fact, according to the World Bank, small and medium sized enterprises (SMEs) represent about 90 percent of businesses and more than 50 percent of employment worldwide, yet approximately half of formal SMEs don’t have access to traditional credit. Excluding this large swath of potentially deserving companies is not good for borrowers, lenders, or ultimately the consumers of the goods and services these companies produce.

The good news is that credit providers are no longer limited to lending practices that have confined capital markets in the past. More and more, lenders are looking to new technologies, like artificial intelligence and machine learning to enhance their decision making. With these tools, capital providers can evaluate vastly more data points in their decision-making. Moreover, with the ever-increasing adoption of Web3, the transparency inherent in blockchain technology will bring to bear an entirely new category of information.

The Consumer Finance Analog

When thinking about commercial financing, looking at the market for consumer loans is instructive. The now-ubiquitous personal FICO score, which was created over thirty years ago, was arguably the first algorithmic credit protocol. This score takes into account payment history, income, and debt-to-credit ratios, among other inputs, and reduces an applicant’s creditworthiness to a single number. In the case of consumer credit, while the FICO score is a convenient shortcut that has streamlined the approval process, these calculations can end up excluding younger individuals or others who have less established credit profiles. In 2020, the average FICO score of young adults 18-29 in the United States was 677, the lowest of all age demographics and 34 points lower than U.S. average of 711. Likewise, some with instances of poor credit behavior in their past might nevertheless be better credit risks than their histories suggest.

Modern protocols and artificial intelligence improve on this process in a couple of ways. First, modern lenders can take into account vastly more than the reportedly less than 50 inputs FICO scores include. Data points from non-traditional sources like utility bills, rental payments, cell phone, and cable bills, social media sites, online search histories, and other “Big Data” give lenders greater insight into whether individuals who might have fallen into the category of the so-called “credit invisibles,” may be more willing and able to pay off loans than the old analysis would suggest. Secondly, and perhaps more interesting, the latest protocols, beyond simply analyzing higher volumes of data, actually have the ability to learn and get better at predicting the credit behavior of the people they extend financing to.

At the enterprise level, capital markets are benefitting from a similar expansion of inputs and learning algorithms that get better at measuring credit risk without human input. Startups and other non-traditional borrowers can have their applications bolstered by the inclusion of receivables and sales histories, customer reviews, and their repayment histories. Industry-wide trends may also be used to better reflect these borrowers’ creditworthiness and improve their prospects. All of this results in closing the gap left by traditional methodologies to extend credit to enterprises that have too often been left behind.

The Competitive Edge of AI

It’s also important to note that, besides providing better opportunities for borrowers, these new technologies have immense benefits for the supply side as well. Lenders who do not incorporate them fail to do so at their own economic peril. In particular, while conventional overemphasizes past behavior in its analysis, artificial intelligence also takes into account current industry trends and future potential. Including these additional inputs leads to more accurate assessment of borrower risk that means less non-performance, more appropriate pricing, and more profits for lenders who use this technology to make faster, cheaper, and more predictive credit determinations. And with more inclusive analytics, these lenders expand their potential target market to include more potentially profitable lending opportunities, filling in a niche that traditional lenders are currently ignoring.

Lastly, AI-driven analytics is a potential service that lenders can offer their customers.

Using AI-driven analytics, they can give startups the ability to analyze their spending and expenditure patterns, providing transparency and insights on how to control their costs. These value-added services thus provide an additional revenue stream for lenders, adding to their profitability while simultaneously improving borrowers’ ability to manage repayment.

Of course, none of this is meant to suggest that AI and credit algorithms are perfect. Especially as they take advantage of machine learning, their protocols can become less connected to their human-guided origins, and their methodologies and biases less transparent, causing a potential “black box” problem. But in the end, with vigilant attention to their results, the positives of this new technology will vastly outweigh these concerns and usher in a future of finance that is more inclusive and efficient on all sides.

The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.