What are Likely Outcomes from Using AI in Banking Applications?

I have been recently thinking about AI…a lot.  I am working on three Generative AI (“GenAI”) services that would improve the ability for FNBB to service its customer FIs and enhance our ability to communicate effectively on strategic issues. AI is mentioned in nearly every industry periodical as the next “thing” and pushing you to get ‘on board.’  But I have very specific service and advocacy outcomes in mind for my current AI projects, and I believe you should also focus on outcomes when investigating rolling out an GenAI product or service.

Here is an illustrative graph from a recent Financial Brand article that shows the types of services bankers expect GenAI to enhance:

This graph is formatted to show the level of bank customer interest on the Y axis and the X axis indicating the level of FI interest.  Thus, the higher and to the right an element is, the more relevant it is to both customers and their institutions. Let’s examine these in that order:

  • Fraud Support – There is no question about the need to identify and remediate fraud. Using advanced AI algorithms, the GenAI tools can establish patterns of known good payments or customer access behavior and immediately detect anomalies. GenAI is powering new check fraud solutions that can examine the imaged versions of checks and identify fraud suspects. GenAI is making inroads in detecting money laundering and other BSA related events.  GenAI is also providing support for identification of potential fraudulent receivers, a big deal as faster payment options ramp up into production.
  • Personalized Offers – The ability to access data related to a customer’s activity can be an important element in a successful “next best product” campaign. Going way beyond knowing whether a customer already had a service to avoid pitching it to them, GenAI creates a model from existing customer profiles that show high levels of profitability to then determine what mix of products and services a customer would need to achieve that same level of profitability.
  • Virtual Assistants – The main virtual assistant we offer customers is online chat, but an effective GenAI powered chat would eliminate the frustration that traditional chat systems cause by understanding the intent of an end-user’s query. Then the GenAI chat would be able to deliver information to the end-user contextually accurate.  Further, if this chat was enabled for online or mobile banking, the GenAI chat could take the user directly to the specific feature that was the target of the query.
  • Voice Assistants – One of my AI projects is the creation of a GenAI created virtual assistant I have named Buckley. Buckley is being trained to offer financial advice targeted to younger bank customers and prospects. It is my contention that these Gen Y and Gen Z individuals will access their phone to receive wisdom from their financial institution.  By having a GenAI powered agent such as Buckley, which has the guardrails in place to limit the information and advice that is proffered, an FI would be in a position to offer financial wisdom in the very manner most likely to be received by younger customers, a strategically important demographic.
  • Wealth Management – Similar to the value of an interactive assistant, specialized digital agents could be trained to make complex assessments of an individual’s financial profile and make recommendations to achieve goals such as saving for college, the purchase of a home or the capital needed to start a new business. Also, the desire for younger customers to access an online agent on their phones would apply to wealth management. Consider that over the next 20 years, trillions of dollars will pass generationally from baby boomers to Gen Y and Z.  Will a 30 something Gen Y come into the bank to make an appointment to discuss wealth management with an advisor more than twice their age?
  • Credit Risk Assessment – GenAI is enabling customized credit modeling so that the uniqueness of certain industry verticals or other such variances can be factored into the data that is presented to officers for loan decisioning. Post loan issuance, GenAI is powering information about the likelihood that a loan customer may go into default or bankrupt. This can make decisions about loan loss reserves and proactive customer interaction timelier and more effective.

These are not the only options where GenAI can power some type of financial service that increases the customer experience, generates more revenue or decreases operational costs. You may have individualized needs where GenAI can become a mission critical tool. My point is to have a specific outcome in mind when you are pursuing an AI powered solution. Don’t fall for AI hype because there is plenty of that out there. When you hold vendors to account to show specifically how their solution will achieve a certain, specific result you have articulated, then you can accurately assess whether that vendors solution will address the specific outcome you desire.  For example, in my research on a GenAI powered online chat, the key element is the ability for the solution to accurately understand intent.  Therefore, you must create multiple real-world examples of what a user might actually key or speak into a query box and then see how well the tool determines the actual intent.  If the vendor is allowed to only show their pre-determined demo queries and outcomes, I have not really tested that solution to determine how my desired outcome(s) would be achieved.

Generative AI is here to stay and every FI should be actively examining how it can best be utilized to achieve its strategic goals.  This starts with brainstorming the most beneficial uses for GenAI for your institution and then examining the commercially available solutions that address that particular outcome.  Look at two or even three vendors that claim to provide that solution.  Make the vendor demo to your outcomes and ask questions about their future development plans.  Things are rapidly changing and every vendor may not have all the of specific elements you need.  Your purchase decision comes down to identifying the vendor that meets all of the required elements and who shows that they are actively planning for the changes that are sure to come in the next 18 to 30 months.

I value the opportunity to chat with you about your plans for using AI powered tools.  Reach out to me at dpeterson@bankers-bank.com and let’s talk about your plans for the evaluation of and deployment of GenAI tools.

 

Resources

https://thefinancialbrand.com/news/data-analytics-banking/artificial-intelligence-banking/catching-up-and-moving-forward-with-genai-in-banking-179666/

 

The views expressed in this blog are for informational purposes. All information shared should be independently evaluated as to its applicability or efficacy.  FNBB does not endorse, recommend or promote any specific service or company that may be named or implied in any blog post.