How often have you called a customer service number and gone through this exchange: “Hello, this is [fill-in-the-blank]. Please say your name. Now say your account number. Now say your PIN or password.” If you get this far, how often did the IVR get it wrong? Finally, in tears, you beg for a real person to help you. Screaming, you demand, “Give me a real person! Give me a customer service agent!” Then either someone picks up the line or you hang up in frustration. This is not AI at work. Not even close to real AI. So, what can AI do for us today?
AI, or artificial intelligence, is described differently by various industry experts. AI covers a wide spectrum of capabilities, from Robotic Process Automation (RPA) to Natural Language Processing (NLP) to Cognitive AI.
AI today is real, but not quite the fully-thinking, understanding, smarter-than-human robot we see in the movies and on TV. Don’t get me wrong – we are moving in that direction. Over the last few years, AI technology is improving and becoming smarter. For example, Natural Language Processing is now beginning to be used, and the robotic “understanding” and “response” has become more highly evolved. Most of us have heard of IBM’s Watson. Watson is an NLP engine that understands most of spoken or written language. Of the many NLP engines in the market, some are uncanny in their accuracy; others are still missing some of the semantics. NLP is one of the primary tools in today’s customer service chatbots. We’ve seen chatbots that can understand different languages, accents, noise conditions, and regional dialects in speech.
Several financial institutions, like Bank of America and USAA, are implementing chatbots that use NLP to communicate between customers and customer service. AI is being used as a customer service representative’s tool (to find the right answer to a customer’s/member’s question) or as a direct responder to customer inquiries (either to different types of customers or outside of normal service hours). We have researched chatbot vendors in the market and of the dozens that work with financial institutions, we’ve seen a wide variety of quality in these chatbot vendors. So, although AI has come a long way over the last few years, don’t take the fact that the quality of chatbots has improved dramatically – to the presumption that all chatbots are working well and smoothly.
Real AI is also being implemented at the enterprise operational level – not just at the customer experience level. Customer experience AI are the chatbots described above (including the virtual assistants and intelligent speakers, like Amazon’s Alexa or Google Home). Enterprise operational AI use cases are tools banks and credit unions have implemented to lead to a better credit or fraud risk model or a more efficient ecosystem to process transactions and make product offers.
Just a few weeks ago, Jack Henry hosted a panel of AI experts at a Real AI session at our Symitar® Educational Conference where we discussed real AI with representatives of Clinc, Faraday, and Infosys – and each highlighted some of the ways AI is being used by their companies and financial institutions today around the world. Clinc’s main use case for AI is its ability to understand customer service questions and in some cases to proactively provide recommendations to the customer based on the financial institution’s product offerings and data availability. Faraday focused its use case example on their ability to better target offerings to prospects or to better identify cross-sell opportunities to their financial institution’s clients. Finally, Infosys discussed several of their many use cases. What struck me as memorable was that many of their use cases were the next generation of solutions developed over the past several decades. The difference was the way the AI engine tied multiple predictive models into a “system” that replicated human thinking. That process could lead to a superior credit evaluation (and expansion) process, a more sophisticated real-time fraud reduction process, a more dynamic cross-selling and product upgrade process, or even a more customer-friendly loan application and onboarding process.
The bottom line is that we are beginning to see AI as an upgrade to many of today’s systems and processes that might not appear to the customer as radical changes or improvements. More likely, they will be perceived as natural progressions to how financial institutions interact with their customers today. Stay tuned as these improvements get dialed into our day-to-day customer experiences – from the top tier financial institutions to our community banks and credit unions.
 Where manual processes are converted to automated processes
 Where most things a person says, or writes can be understood and responded to
 Where companies use predictive analytics and other tools to understand the customer, and respond or recommend a course of action unique to that customer