by Robin Austin

Are you happy with chatbots replacing humans?

Jul 09, 20197 mins
Artificial IntelligenceFinancial Services IndustryMachine Learning

AI and machine learning are predicted to be the most positive disrupting technology for banking with an estimated cost saving of more than $1 trillion by 2030. Although AI/ML fits into the 'do more with less' mentality and the hope is that chatbot functions will personalize service meeting customer needs faster and more efficiently, it has fallen short so far.

chatbot chatbots customer service helpdesk automation by oruj hummatov getty
Credit: Oruj Hummatov / Getty Images

At least 95% of business professionals according to Forester have stated that utilizing chatbots to gather customer transactional history through live customer interactions is very important. Customer-facing tools like websites have been gathering and tracking customer traffic and performing search engine optimization (SEO) to stay “top of mind” for many years.

The hope is that chatbots will gather immediate real time “buying” habits supplying data to a backend analytics engine that can offer suggestions and increase sales. Ultimate, the view is that a customer may answer chatbots more easily than real customer service people when requested for personal data. Is talking to a personal machine easier then to a person? Many customers have said that they answer without thinking because it is less personal talking to a machine. We opt-in without thinking. Why not chat with a bot without thinking?

Chatbots are permeating our lives behind the scenes

Currently, chatbots are fairly limited and do a great job of answering pre-defined Frequently Asked Questions (FAQs).  The hope is that chatbots can respond in a personal way by understanding personal context through conversational dialog and respond in kind. Beyond that, chatbots should immediately identify the name and any opt-in information without having to ask. As soon as the customer opens the door to the website, the chatbot should be able to reach out in conversation following, watching, asking questions and offering additional information as they traverse the website. This is somewhat successful but for many financial institution executives not yet as capable as they would like. The same view is shared by their customers.

The view of many internal company marketing and sales and customer service personnel is that third-party chatbot artificial intelligence (AI)/machine learning (ML) software tool providers are DevOps people without the understanding of the necessary subtleties and interactions. Chatbots need to emulate a person seamlessly, tapping into an AI/ML system with the necessary behavioral characteristics captured along with having comprehensive over all company and product knowledge. Right now, due to a third party’s inability to comprehend this personal attachment, tone and language subtleties, not enough has been done to truly reflect the company’s brand, reputation and communicate the value of their products without it sounding sterile or hyped. They believe that many DevOps people don’t yet have the expertise on staff necessary to personalize programmable functions.

Banks rush to implement chatbots

More global Fintech start-ups and full-service banking institutions are rushing to implement chatbots for customers to manage their money. Customers are all about experiences when they choose their bank. Today, it is a personal choice. Digital services offered on mobile devices that help them manage their banking lives and beyond is important to them. Real-time smart seamless digital services delivered in a split second when transferring money, shopping and paying bills can improve or damage an institutions reputation. Seamlessness with a superior real-time experience is required for financial institutions to keep loyal customers and maintain relevance. 

Individual brands are creating chatbots on social networks, in addition to chatbots on their own websites and mobile apps. Giants like Google, Amazon and PayPal are making their way into people’s financial lives in the US.  By offering digital-first platforms, supporting Omni-channel banking, smart banking, modular banking and open banking, banks have major competition from companies never thought of as competitors in the past. Banks can compete with the digital giants, but they need to move fast or they will be replaced.

French bank Societe Generale is developing chatbots to answer FAQ like equity fund questions in a Romanian banking unit. TransferWise has partnered with Facebook providing chatbot services that allows customers to send money to friends and family internationally though Facebook Massager.  Swedbank’s “Nina” chatbot has received and fielded over 40K conversations a month. Swedish banks have implemented AI/ML chatbot platforms to improve customer satisfaction because customer satisfaction is at an all-time low.

“Trust” has dwindled due to poor customer service and increased banking costs. Although many believe that by implementing AI/ML chatbots customer satisfaction will improve, those financial institutions positioning themselves ahead of the curve have only seen early adopters feel comfortable and are happy with chatbots for their customer service help. Swedbank has implemented AI/ML for customer service over all customers and only a select market segment shows improved customer satisfaction when surveyed.

Most global banks now wished they hadn’t been so exuberant to implement chatbots full board across all services and customers. It seems customers have very low expectations for self-serve customer service when the full human service has failed. Most customers believe that chatbots may be just an added layer with no results. They believe that more energy will be expended, and their valuable time wasted still having to call the support line and start again. Unless the chatbot can emulate and improve on a full customer service personal experience, the customer feels they are subjected to an extra step that doesn’t solve their problem.

Meet Erica, Eno and Zelle

Bank of America (BofA) calls their newly developed chatbot “Erica” believed to have been taken out of the word AmErica. Erica helps customers perform simple transactions like reporting credit card amounts and due dates, performing checking and savings transactions like transfers, deposits and balances and offering suggestions on paying down debt. Through an ongoing behind the scenes real-time AI/ML behavioral spending analytics assessment, Erica can determine the buying habits of the customer and offer better spending suggestions to avoid too much debt.

Along with Erica, BofA uses Zelle, a peer-to-peer money transfer service. Some of the largest banks in the world — Bank of America, Chase, Capital One and USAA have Zelle as an offering for online banking also available through their mobile app. Zelle is an instantaneous payment processing network easily completing transfers instantly to other users. Even if one of the participants is not yet using Zelle, money can be sent through phone or email allowing the user to download the Zelle app to receive their money.

One of the few gender neutral chatbots is Eno from Capital One. Eno sends text messages notification reminders to make credit-card payments on their mobile phone. Eno also passes along text messages with security alerts and any verifying transactional activity.

In my discussions with Financial institution executives, most have begun to seriously consider AI/ML solutions like chatbots very customer friendly and a possible valuable alternative to high cost customer service personnel. Whether customer service will be improved is not really the most important to them. They believe that chatbots will reduce their costs affectively and provide a “more than satisfactory” alternative at this point in time. They are gambling on chatbots’ future and the speed AI/ML systems are improving. Customers will put up with temporary weaknesses until one day they forget that chatbots are not human.

The use of chatbots, in the banking executive’s view, improves their image as an innovative technology company and this image far exceeds any current chatbot glitches or lag time in development. More than one banking and finance exec has said to me, “We will get there very soon and just like cell phones we won’t remember or be able to live without them.”