Among the biggest stories in tech remains the astonishing advancement in artificial intelligence, or AI, over the past several months. While AI has been evolving for a very long time, its latest iterations and implementations have reshaped how people think about AI and its capabilities.
Much of the conversation around AI has focused on AI-generated art and ChatGPT, both of which are able to accurately follow detailed prompts and create reasonably convincing works. And many industries have already begun utilizing AI to aid their businesses. What does all this mean for financial institutions? Finally, how can bankers leverage this innovative technology to help their organizations and their end users?
AI has many potential use cases for financial institutions and their processes. One of the most intriguing, in light of ChatGPT, is the ability to interface with customers in a very personalized way. Banks can use large language models, like GPT-3, to automate customer service. AI can take instruction and communicate back with an operator towards a specific goal in a way that feels human and can be very helpful. AI is also capable of following very specific communication styles and guidelines.
This gives banks a way to implement AI that enhances the customer experience. A bank customer with a question or seeking an update on a loan can ask an AI-enabled chatbot that can check the loan’s status on the back end and effectively communicate with the customer. If the bank has customer contact information, the AI can also automatically reach out to a customer and request whatever documents or information the bank needs to complete a loan or any other transaction.
AI can also help in a bank’s back office. AI’s ability to process massive amounts of information, coupled with its ability to change and improve over time, makes it a compelling candidate for decision making roles. One example already in operation is the use of AI in loan underwriting as an alternative to one-size-fits-all credit scoring models like FICO. Once the financial institution has gathered the information needed, an AI can recommend a decision to either approve or deny a loan using parameters that represent the customer’s unique demographic and provide a fairer method to evaluate risk. Even if that AI cannot make a decision on all pending loans or come to a complete decision, it can do the heavy lifting and eliminate the bulk of manual work for bank employees.
Another advanced role AI can play for financial institutions is known as optical character recognition, or OCR. OCR can extract text, handwriting and data from images and documents and move it into a system of record without manually inputting the information. Operators can automatically scan images or PDF files of document as they arrive and automatically draw all the information from each of them. Bankers can use OCR to radically increase efficiency for a wide range of document-based processes.
However, there are inherent caveats and shortcomings in all the aforementioned AI banking use cases. AI decision making is not perfect and requires human supervision. In fact, overreliance on AI for tasks like underwriting can lead to biases and ethical oversights. Even its use as a chatbot is not as straightforward as it sounds. ChatGPT, with all of its recent buzz, is far from perfect. Like most AI solutions, it cannot yet handle highly detailed financial information. Banks looking to implement AI should seek solutions that are specially tailored for financial services, which provide a higher degree of accuracy thanks to its specialization.
Finally, it should be noted that there is no magic bullet solution or product for any financial institution — including AI. Hype is a powerful force in both technology and finance, and it is easy to get swept up in the excitement. It is vital that any organization looking to invest in AI technology start with mapping and deeply understanding their own existing processes and use that starting point to determine the areas best suited for automation enhancement. The right way to think about AI within banking is to understand that while it can’t automate every task, it can eliminate large parts of the manual repetitive work that may be slowing down your financial institution today. That, for any team, is transformational.