ChatGPT, as we all know, went from nonexistent to fodder for dinner-party conversations and late-night monologues in the span of about two weeks—an unprecedented hype curve. Many of us are supposedly going to lose our jobs when they are automated away by this technology.
Well, here at Aite-Novarica Group, we decided to test these notions. Can an AI write as well as or better than a human? When AIs write for their human readers, are they able to properly consider nuances of the human condition? Can an AI achieve insights and write about them better than a human can?
We want you to help us decide.
Below are two blog posts written about the expected impact of ChatGPT on commercial lending: one written by Chat GPT, another by Aite-Novarica Group Strategic Advisor David O’Connell. Have fun reading and comparing them, then use the comments section below to let us know which has more value for a manager of a commercial lending line of business.
AI vs. Advisor
Artificial intelligence (AI) has the potential to revolutionize the commercial lending industry. One of the most significant impacts of AI, specifically chatbots like ChatGPT, will be on the customer experience. Chatbots can provide 24/7 customer service and assist with tasks such as loan application processing, account management, and loan servicing. This will not only improve customer satisfaction but also reduce the cost of operations for lenders.
Another impact of AI in commercial lending will be on the underwriting process. AI-powered systems can analyze large amounts of data quickly and accurately, making the underwriting process faster and more efficient. This will also help to reduce the risk of loan defaults by identifying potential red flags in loan applications. Additionally, AI can assist with fraud detection and prevention, further improving the security of the lending process.
AI can also help commercial lenders to better target and segment their customer base. By analyzing customer data, AI systems can identify patterns and trends that can be used to create customized loan products and marketing campaigns. This will help lenders to reach new customers and increase loan origination.
One example of the practical application of AI in commercial lending is the use of natural language processing (NLP) technology in loan application processing. NLP can be used to extract information from loan applications, such as income and credit score, and input it into a computer system automatically. This can significantly speed up the loan application process and reduce the need for manual data entry.
In conclusion, the implementation of AI in the commercial lending industry has the potential to greatly improve the customer experience, increase efficiency, and reduce risk. As technology continues to evolve, we can expect to see even more advancements in the use of AI in commercial lending in the future.
(Here’s the originating prompt: Write a 250-word blog article about the biggest impacts that ChatGPT and other forms of artificial intelligence will have on different aspects of commercial lending. Make this article to the point, not embellished, not academic, in the tone of an analytical professional industry research analyst, in neutral third person, future tense, predictive of the future, detailed real-world examples, professional level writing.)
Commercial lending operations that embrace ChatGPT will reap huge benefits. In fact, for early adopters, ChatGPT is the right technology at the right time. By automating away tasks that are painfully necessary but also costly, labor-intensive, and additive of little value, ChatGPT will enable commercial lending operations to opt out of the war for talent, increase productivity, shift personnel to more important tasks, and make desired growth goals more achievable.
ChatGPT will confer the most benefits on loan officers and underwriters by improving their productivity and optimizing the mix of tasks they focus on. Currently, loan officers and underwriters spend lots of time on borrower financial statements: obtaining them, “spreading” them so that they are consumable internally, analyzing them, identifying the stories they tell, and writing about what those stories mean for the borrower’s creditworthiness. This set of tasks, although essentially analytical, is full of rather rote stuff. Writing, which comes easy to precious few of us, is labor-intensive and costly. Although a bit more intellectual, borrower financial analysis is essentially the crunching of numbers for comparison to loan terms, historic financial performance, and the borrower’s loan covenants, whether proposed or existing. If artificial intelligence (AI) tools can spread financial statements—they can—and ChatGPT can write content such as journalism pieces and even articles about itself, then certainly it will be easy for this capability to do any of the following for any particular lending arrangement: consider a borrower’s financial spreads, peruse loan-officer call notes, divine the story, correlate that story with the a borrower’s debt-repayment commitments, and write some of the most important portions of credit proposals that get reviewed in credit-approval processes.
Far more important about ChatGPT’s ability to automate away commercial lending’s most costly and rote tasks is its ability to enable growth. When I ask folks who manage commercial lending lines of business how fast they seek to grow their business, they typically have two answers: the growth rate that is permissible or sought by senior management, and the growth rate they’ll actually achieve, with their best efforts severely curtailed by the war for talent. Commercial lending has it particularly tough in the war for talent. New college graduates don’t want to go into banking; they want to work at fintech companies, social networking companies, or companies that invent new AI tools such as ChatGPT. But guess what? Maslow’s hierarchy of needs does not apply to ChatGPT. So it’s perfectly fine doing the supposedly uncool work of commercial loan underwriting and doesn’t expect its job to result in instant self-actualization. Even better, ChatGPT doesn’t require costly and time-consuming training, be it on the job or in a formal credit training program. Boom! Growth impediment vaporized.
Also good about ChatGPT is what lenders and underwriters will do with their newfound productivity. It won’t be spent browsing the news or Facebook; that’s not how this driven crowd rolls, especially at institutions with well-structured benefit plans. Unburdened by writing and analysis, underwriters and lenders will spend more time on the bullet points on their job descriptions that are most specialized and require human intellect. Among these are coming up with deal structures that will simultaneously be accepted by the borrower, deter the competition, and be amenable to credit-approval authorities. Another area that will get more attention is projections. With less time dedicated to rote historical analysis and writing, underwriters can spend more time on the creative, risk-mitigating, and iterative process of tampering with a borrower’s pro forma revenue, profitability level, and cash flow, to see how the borrower’s creditworthiness will hold up under different potential scenarios.
Lenders, for their part, will spend more time on the activities that drive their organization’s risk-adjusted return on capital. Among these: cross-selling noncredit services and ensuring that the onboarding and delivery of these services, once sold, delight customers. And let’s not forget about moving on to the perpetual next deal. The less time they spend getting approval for each of their new deals, the more new deals they can field. Growth indeed.
And about lenders and underwriters spending more time with borrowers: This is where ChatGPT can’t go, and it’s exactly where business-lending financial institutions need their people spending more time. Because ChatGPT might be able to surf data, deduce insights, write about what it discovers, and do it with incredible speed, but there are plenty of things it can’t do. Humans, we sometimes forget, are spectacular pattern-recognition machines, especially in human-to-human interactions. Humans, not ChatGPT, can gauge the enthusiasm in a prospect’s face when cross-sold something that will make their controller’s department more productive, or notice when a chief financial officer seems uncharacteristically unsteady. What might be their situation? Under the weather? Affected by our burgeoning mental health crisis? Trying to conceal financial problems? ChatGPT will actually enable seasoned lenders more time and latitude to establish with a borrower the connection and trust required to be a valued advisor, which is the role required for advising, lending, cross-selling, and risk monitoring on behalf of the business-lending financial institution.
And all that fear and loathing about ChatGPT currently in the zeitgeist? It will likely go the way of the resistance that was held for now broadly accepted systems such as customer relationship management and commercial loan origination. Sure, people might initially fear being laid off by a technology and resist the change it represents, but business history is rife with technologies that, though initially feared and sometimes reviled, ultimately made knowledge workers’ lives better by eliminating the mundane, leaving more time for the challenging, and ultimately improving productivity. And wherever there’s productivity, there tends to be job security.
Don’t forget to leave us a comment and let us know who wins the battle of AI vs. Advisor.