Technology is transforming the financial services industry in ways that would have been unimaginable just a few years ago — a trend confirmed by Capital One’s 2019 survey of the asset-backed securities (ABS) professionals at SFIG Vegas. Ninety percent of the respondents declared that adopting new technologies is a critical component of their preparation for the next economic cycle.
The single most important of these new technologies, identified by 35% of the respondents as the primary driver of change, is artificial intelligence (AI). Although artificial intelligence is not new, it increasingly has the power to transform, thanks to a confluence of developments. The first is AI’s growing sophistication and power.
The emergence of AI as a practical tool has also been fed by dramatic advances in the computing power needed to run complex algorithms, by greater access to this power through cloud services and by the ready availability of vast amounts of data that AI requires to perceive patterns and make predictions. With increased ability to capture, digitize and aggregate information, whether from public records, credit card transactions, social media posts or online searches, we have reached a point where there is sufficient data to help progress even the most obscure AI challenges.
AI is being used to diagnose diseases, translate languages and guide autonomous vehicles — and many of the algorithms behind these tools are being applied to the financial sector. Lenders are already using them to automate and streamline documentation, identify new markets and even assist in managing capital markets activities.
Streamlining the Approval Process
The pattern recognition inherent in artificial intelligence and the analytical power of the smart algorithms that underlie it offer many benefits for lenders. Among them, the ability to automate and streamline an increasing number of aspects of the approval process. While this advantage is more pertinent to consumer lending, where standardized forms are the rule, it is being adapted to the commercial finance approval process as well.
For instance, AI can easily learn the types of acceptable answers for each area of a loan document, conduct a preliminary review and flag exceptions for further examination. Not only does this application of AI reduce processing time, but it enables underwriters to focus on higher-level tasks.
Furthermore, the ability of AI-powered programs to understand how rules are to be applied in different situations helps lenders make sure that the information they require of borrowers and the processes they use to evaluate them comply with relevant local laws. This is particularly important for national lenders such as mortgage finance companies that must adhere to overlapping federal, state and sometimes local rules.
Reaching Out to Likely Borrowers
Lenders are using AI to identify likely borrowers for their loans, pinpointing specific companies and leaders who are most likely to be interested in financing for a particular purpose — to purchase equipment or secure a line of credit, for instance — and who are also good credit risks. Essentially, AI is helping lenders make the most of their funds by allowing them to narrow their focus to high-potential customers and tailor their products accordingly. In many cases, lenders are even using AI to help identify the right traditional or social media channel to reach each target group.
New Road to Securitization
AI’s emerging role in securitization has made international headlines. Fintech startup Pagaya Investments announced this year that it has privately placed a consumer-loan securitization that will rely solely on the big data analytics of its proprietary AI platform rather than human judgement to manage and operate a $100 million portfolio. In fact, Pagaya used AI to select the unsecured online-loan assets in the first place. The company’s CEO, Gal Krubiner, believes that AI will soon direct decisions for CLO, RMBS and other ABS-asset-class portfolios.
A New Approach to Underwriting?
The influence of AI is being seen in large ways and small. Perhaps, most controversially, it is starting to be used to assess the credit risks of consumer and commercial loans. Traditionally, underwriting has largely relied on a series of well-established financial calculations to determine a business’ or consumer’s capacity to repay a loan. AI represents a fundamentally different approach. The conclusions it draws might be derived from comparing an individual business or consumer with a cohort of hundreds of thousands or even millions of businesses or consumers with its exact characteristics. Given a specific collection of data points and a large enough data set of comparable borrowers, artificial intelligence may be able to deliver a highly reliable assessment of creditworthiness that lenders can use in combination with or in place of traditional methods.
However, this usage is already creating transparency questions and will continue to face headwinds as regulators seek to understand companies’ complex algorithms.
In addition to increased accuracy, this broader approach may allow a more comprehensive lending decision to be made that might, for example, include willingness to pay as well as ability to pay. It also has the potential to allow lenders greater leeway in fine-tuning their pricing on a case-by-case basis. Finally, the inclusion of these nontraditional metrics also make AI particularly valuable for assessing the creditworthiness of start-ups and businesses who do not have a significant credit history. Essentially, AI is allowing lenders to potentially access a market that had been out-of-reach before.
AI offers comparable advantages to commercial lenders. Factoring companies, for instance, are starting to use AI to better assess the value of the receivables they finance as well as to accelerate the decision-making process, which is crucial in such short-term transactions. AI has also started to make its appearance in equipment finance, where its ability to harness nontraditional forms of information to determine creditworthiness overcomes one of the challenges that equipment finance companies confront — the limited amount of financial data generated by their small business borrowers.
Extending the Business Cycle
The emergence of AI is also having a more macro-scale impact. It’s possible that one reason that the current business cycle has lasted as long as it has is that technologies like AI have enabled lenders to find value and secure competitive advantage that couldn’t be found before, whether it’s making better credit decisions, improving business operations or enhancing the customer experience. As a result, one could argue that more credit has flowed into the economy because lenders have been able to serve more people with less risk. •