Artificial Intelligence (AI) is rapidly transforming both the technology and credit union industries. AI is increasingly being used in innovative ways, enabling new credit union successes. The AI technology landscape is rapidly changing, motivated both by technology capability and business advantage. With these rapid changes, it can be challenging to find well-fitting opportunities to use AI while also considering the risks associated with it.
AI opportunities
The most common services that credit unions offer members are based on financial transactions. These are typically large volume and high-velocity transactions that have a minimal chance of mathematical error. These services require systems to make decisions quickly and are based on pre-decided and programmed “if-then” rules. Any decisions that required learned skills, interpretative analysis, or personal experience, can not be programmed due to the complexity and unpredictable nature of the rule sets. With the arrival of AI technologies, now they can. Any member service or product that you offer today that has complex or unpredictable decisions could be accelerated or enhanced using AI.
Members have a variety of touchpoints with credit unions. Many of these are in digital form including websites, banking portals, mobile apps, and ATMs. While these digital experiences have continued to mature, and become more personalized, they are also limited by programmed rules. AI technologies learn how your members interact and can offer a more personalized response without pre-programmed rules.
Credit unions have implemented AI-based chatbots that answer member questions, assist members with banking transactions and recommend financial guidance. These chatbots empower member service representatives to accomplish more while also virtualizing your credit union subject matter experts (SME) with the chat mechanism as their interaction channel. This is an early form of an AI virtual SME, and we can expect many more of these. Some member service teams are using these AI virtual SMEs to assist them internally to query specifics about the products available for your members. They are essentially digital assistants to help them serve members with improved quality and speed.
One of the fastest and easiest ways to integrate AI into your credit union is through the deployment of Large Language Model (LLM) technologies. Some common names for these services include ChatGPT and Microsoft Copilot 365. LLM technologies consume content available publicly on the Internet and within a credit union’s environment to generate answers in response to questions related to subjects the models have been trained on. LLM technologies can provide insight and even answer questions with adjusted answers based on how the question is asked. Many of these LLM technologies can be licensed for secure use with company internal documentation and limited access for employees only.
With all the opportunities available with AI technology, there are also some risks affiliated with its use.
Potential risks with the use of AI technology
AI is based on technologies that learn and interpret, so answers to the same question can vary. Answers may be only partially correct or even entirely incorrect. In addition, poor training data can create answers that can be biased, and the only way to discover this bias, is through thorough question-to-response testing. This can be challenging since answers can unpredictably vary, so higher testing volumes are necessary. These challenges point to a greater concern that AI answers may not be trustworthy enough to use in critical decision making or to convey information of high importance. Some AI technologies do not have effective adjustment or tuning mechanisms to eliminate trust concerns.
AI systems can be used to help protect against cyber threats and data breaches, but they are also targets for attacks and information leakage. In many cases, employees are sharing information with external LLM’s like ChatGPT and are unaware that their private questions could be used for the LLM training and further public use. It is important that credit unions have policies in place to help protect against these concerns.
AI has arrived, is your credit union ready to implement it?
Artificial Intelligence is here, and we can expect transformation in significant ways over the next two to five years. AI technologies are arriving that are easier, cheaper and faster to deploy. Buy-in is faster in recent years and the entry point for AI technical know-how is lower than ever. Catalyst continues to look for opportunities to leverage AI to optimize our processes and provide value added services to our members.
AI technologies are ready to transform the credit union industry. Is your credit union ready?