“We’re a technology company that happens to do biology.”
— Stephane Bancel, CEO, Moderna.

While lenders and financial institutions are not looking to discover the next vaccine or medical breakthrough, artificial intelligence (AI) and machine learning (ML) have been reshaping traditional business processes at a rapid speed. Many decision makers in the C-Suites, while still responsible for ensuring everyday “blocking and tackling” is done, are dealing with how best to integrate AI/ML into their companies, and at what cost. This article takes a deeper dive into how AI and ML are being utilized in the industry along with the factors leaders must consider with AI/ML integration.

AI and ML are not synonymous. ML should be thought of as a subset of AI, although ML is by far the largest subset, followed by others such as Deep Learning. Artificial Intelligence covers the broad landscape of creating intelligent machines while ML can learn from these creations and continuously improve performance when correctly implemented. As an example, the intelligence of the AI machine of a credit underwriting scoring model will provide a decision or outcome; however, through ML, which runs and depends on receiving continuous, updated inflows of data, the credit scoring system can be “learned” and smartly evolve or to improve the model and further benefit a company.

Not surprisingly, in the financial world, AI/ML was first used in the consumer finance space where troves of data are readily available. Areas such as credit card fraud detection and chat-box services, which many of us use daily, have grown quickly and are continuously enhanced as these models learn and train on data literally received every second of each day. Movement from the consumer financial channels to the commercial channels hasn’t taken long as the benefits of AI/ ML become more mainstream.

Philip Armstrong of Mo Technologies, a credit as a service (CaaS) fintech with proprietary technology in the ML and AI space, has seen the impact of AI/ ML in both the consumer and commercial financial sectors. He gave the example that in Latin America, over 2 billion credit card transactions are declined annually due to non-sufficient funds (NSF). “The data showed that the average NSF was just $6 or 10% of the $60 attempted spend; this represented over $120 billion of gross dollar volume left on the table. Through appropriate scoring, we could determine that 40 – 50% of these should be processed because the default risk is extremely low,” Armstrong stated. “This allows for improved overall customer and brand loyalty with the credit card company, not to mention capturing future revenues that were not available prior to the implementation of the scoring model.”

In the commercial lending space, Mo Technologies used its ML capabilities to help credit-score smaller farming businesses that agricultural banks could not underwrite using traditional credit models. Armstrong cites that “the part of the model that is dynamic is the feedback loop, which constantly refines the model with appropriate frequency.” Working with the agricultural lender, Mo Technologies could capture and train (and retrain) its model for the lender, showing default rates much lower than originally thought. This ultimately opened a whole new sector of borrowers for the bank and access to capital to farmers.

This new access to information through AI is rapidly moving through all sectors in the financing world largely for two simple reasons:1) Low cost and 2) It works. According to a recent Gartner Inc. forecast, AI will be involved in 75% of venture capital decisions by 2025, up from less than 5% today. Take the case of Correlation Ventures, a San Francisco-based early-stage VC firm with over $360 million under management. The firm internally developed a machine-learning tool that reviews information extracted by humans and pitch decks and other materials from startups to decide whether the firm should invest in a company. This information is then fed into an algorithm trained on data from more than 100,000 venture financing rounds according, to managing director David Coats of Correlation Ventures. But AI/ML projects do not churn out overnight or the next fiscal quarter, but rather the investment into ML must have a long-term outlook with support and understanding by C-Level leadership.

Daniel Faggella, the CEO and head of research at Emerj, a leading AI research and advisory company, puts heavy emphasis on this long-term outlook objective. “Businesses should avoid AI toys,” states Faggella, meaning that “toy” applications are technologies or projects taken on because they use AI, not because they solve a business problem. Companies need to be careful not to just get into AI for the sake of getting into AI. Faggella believes that it is critical for companies to map out AI projects and to begin with the end in mind. Expanded further, “Leadership must understand that AI is an investment in a new paradigm of skills and resources, so not all experiments will prove immediately valuable in financial terms; initiatives that build towards core strengths and skills are ultimately valuable to creating transformation, not just surface-level solutions.”

Going back to CEO Stephane Bancel’s quote of calling Moderna a technology company that happens to do biology, illustrates a paradigm shift that took less than a decade to complete. Less than ten years ago, a 20-person Moderna firm was still manually entering nucleotide sequences into Excel spreadsheets per the standard industry practice. Today, because of Bancel shifting Moderna into full digitalization, it has become one of the leaders in the coronavirus vaccine development.

In summary, AI/ML, along with deep learning networks, will serve to benefit the lending spaces not only in efficiency areas, but also in creating opportunities that might not even exist today. The ability to apply AI methodology to multiple segments within a corporation and continue to apply year after year and project after project are where the real benefits of the shift into AI are made.