Think about the most common reason you find yourself reaching out to customer service in banking: it\u2019s rarely good news. Much of the time it\u2019s to resolve a problem, and you\u2019re likely not having a good customer experience. As much hype as there is on how traditional banks are \u201cbehind the times\u201d in implementing technology to improve customer experience, I don\u2019t think full-on artificial intelligence-powered chatbots are necessarily the answer. At least not right now.\n\u2018Contact us'\nHere\u2019s an example: Recently I received a text from a fintech P2P player regarding a password change I had not requested. I immediately was alarmed, since that is a red flag that someone is trying hack my account, and this account is directly linked to my Chase checking account. I stopped what I was doing and immediately tried to contact the P2P provider to advise them of the issue.\nAfter a lot of time navigating the app, I was able to send an email on the subject. But I immediately received a message that my email would be returned in 48 to 72 hours, which was not very comforting. I struggled to find a phone number and once I dialed, there was no prompt for \u201cpossible fraud,\u201d so the system kept disconnecting me. During this 15 minutes of angst, I was not having a good \u201ccustomer experience.\u201d In fact, I was getting really angry and frustrated.\u00a0\nLogical reasoning kicked in and I realized that my best mode of resolution would be to contact Chase.\u00a0 I called and got a person on the line immediately. After explaining the problem once, the representative suggested we immediately delink the account from the P2P fintech for 30 days. She even offered to do so at NO charge, since I was experiencing a potential hacking attack. I was so relieved that I could go on with my day, with no further actions needed.\u00a0\nThat\u2019s a positive customer experience.\u00a0 Coincidentally, the Chase representative didn\u2019t try to cross-sell me Zelle, which is the Chase P2P product. That\u2019s where AI could have stepped in after this call resolved the issue, but it didn\u2019t.\u00a0 That\u2019s a missed marketing opportunity.\nAI as a retention\/marketing tool\nThere are AI solutions which can provide the linkage between all customer interactions: call center, social, and branch. They can flag a customer for potential attrition risk or to cross-sell highly relevant products and services related to the issue(s) the customer is having with the bank. This has to be done very carefully, though.\u00a0\nProblem resolution needs to be done \u201cneatly.\u201d What do I mean? If an AI retention tool alerts you that a customer has made multiple attempts to resolve an issue with the bank for an extended period of time, the response must be human and high-touch \u2013 not more AI or marketing. There must be line where AI has done its job and the humans take the issue to the finish line.\nDepending on the account type, a phone call or an email addressing the situation and seeking to resolve it should be the next step in the process. Once resolution is achieved, the marketing opportunity should be evaluated as in my experience with Chase and Zelle.\nThe behavioral intelligence embedded within AI can certainly be used to guide customers through your app or your site. However, it cannot be relied upon to solve urgent problems or quell the anxiety of an agitated customer. I see it more as a data gathering, analyzing, and filtering proposition than a solution to replace all humans. To satisfy customers who feel their needs are not being met, you need something humans possess and bots don\u2019t: creativity and empathy.\nEvery resolution to a customer experience challenge needs to strike the right balance between technology and humanity. In the end, a bank that cannot choose the right method to answer a question or help a customer resolve a problem faster will find that technology can backfire.\nFintechs will continue to rely solely on technology to limit the overhead costs of human customer care. But if their technology gains mass adoption, they will need to consider the challenges and limitations of being too technology dependent.\u00a0\nThe future of work, automation anxiety and industry 4.0\nOne of my favorite thinkers on the subject of humans vs. machines is Erik Brynjolfsson, a professor at MIT\u2019s Sloan School of Management and director of the MIT Initiative on the Digital Economy and the MIT Center for Digital Business. In 2013, Erik released a TED Talk video on \u201cThe Key to growth? Race with the machines.\u201d I think of his message often when considering the future of humans and technology.\nA lot of the hype around technology replacing people comes from automation anxiety, a fear that robots and technology will replace people. This idea has been circulating for over a century, ever since the Industrial Revolution. Now that we\u2019re in the midst of Industry 4.0, or the Fourth Industrial Revolution, we can expect the same kind of unfounded hysteria \u2013 and potential miscalculation around the value and appropriate use of technologies like AI. This short video, \u201cWhy the Rise of Robots Won\u2019t Mean the End of Work,\u201d does a nice job of encapsulating the ideas of futurists and economists on the issue.\nA recent article by Deloitte, \u201cAI, Robotics, and Automation,\u201d sums it up nicely: \u201c\u200bAs AI and other advanced technologies permeate the workplace, skills such as critical thinking, creativity, and problem-solving gain in importance. Leading companies are recognizing that these technologies are most effective when they complement humans, not replace them\u201d.\nIndustry lines are blurring, and consumers want to have the same frictionless experience with their banks that they have with their social media and retail relationships. Yet nothing is more personal than finances and financial well-being. Technology can contribute limitless value to humans, thanks to the power of data and analytics to more efficiently serve banking customers. But banks must never forget that the person is still the pivotal component of that relationship.