I would like to hop up on my soapbox to talk about Artificial Intelligence (AI) and its application in community banking. AI is emerging very rapidly into our everyday personal and business lives and it doesn’t seem to be slowing down anytime soon.
Our receiving or giving of gifts of devices such as the Amazon Echo or a Google Home is helping to spread the exposure and integration of AI devices in our everyday lives. This is likely to sound strange coming from a software guy, but I am not really fond of the thought of having these devices in my home, at least not yet, until more is known of the data privacy and its use by these devices and the companies behind them.
Just as I was slow to adopt Facebook because of concerns 8 years ago that my usage, habits, and data generated was being used in the background to sell me products. With the recent revelations of FB providing data improperly to a large data firm, my initial concerns over data use and privacy have come to realization.
But let’s set data privacy concerns aside for now and focus on the real issue I have with AI. I always thought of myself more of a banker than a software guy, which means I believe we bankers use common sense when thinking about risks and new technologies. But lately I am starting to call that into question when I see what is happening in the banking industry I serve. I believe the adoption of AI into community banking and its application for commercial lending should be done so only after very careful evaluation.
There is no shortage of local, regional or national banking seminars and workshops with AI applications as part of the agenda. The opportunities to learn about AI and its impact or utilization on community banking is plentiful at this time. My caution to bankers and their IT departments is to tread carefully. I have made a living off of bringing software technology and it’s efficiency to community banks and credit unions for the last 20 plus years. There are more efficiency gains to be squeezed from our systems. However, all efficiency gains we have advocated over the years are done so with a focus on supporting bank or loan growth while holding overall overhead costs constant as a result of efficiency gains.
On the other hand, AI and its applications have the potential impact of eliminating jobs and people across all divisions. It literally is robots or robot-like applications taking over for humans and their jobs. While this is important in phases of repetitive manufacturing and such, do you really believe this is good for community banks and community credit unions that differentiate themselves based on personal relationships and personal service? I am not saying AI does not have its place in community banking, but it should be balanced in order to avoid sacrificing delivery of personal service and building personal relationships with your customers.
What about AI in specific areas of lending? For mortgage, consumer and possibly small business loans, it has some very real application, because it can be repetitive in nature, following a based set of rules, policies and conditions.
For C&I, CRE and AG lending, I can not advocate, at this time, that these types of loans be underwritten, approved, and serviced through total AI applications because of the large variation in conditions across the universe of business, commercial and agricultural lending. The reality is, it’s 2018, and I am still prodding the vast majority of community banks and credit unions to move their commercial and agricultural lending off of Excel, Word and other non-integrated and disparate systems that reek of inefficiency and lack of data control.
When I see institutions such as these suddenly want to leap frog to full automation, I scratch my head. This leads to turning over their customer and staff to AI involving auto analysis, auto decisions, and auto approvals and in the process eliminates intervention of the credit analyst and loan officer or loan committee. I would advise proceeding with extreme caution. Again, I am 100% behind the process of improving efficiency in lending in community banks, but I feel strongly we need to keep qualified credit analyst and loan officers in the equation and use moderation and common sense when adopting AI and automation in your institution, particularly in the commercial or agricultural lending areas of your operation.