How banks can choose the right chatbot

They may have been a long time in coming, but banking chatbots are coming to a sofa near you soon! That’s the message to consumers from financial institutions. Chat Platforms and Bots are set to replace apps and browsers and make personal banking a whole lot easier. It’s like going to a branch and chatting to a real human being, but without the effort of having to leave your sofa.

Bank of America has just introduced a new chatbot banker to the world. Her name is Erica. MasterCard’s new chatbot banker is named Kai. Erica and Kai are the culmination of investments that the two companies have made in smart messaging technologies designed to answer account questions, offer financial advice and handle some transactions. But there’s quite a bit of debate on whether chatbot bankers are ready for prime time.

Recently, Peter Wannemacher of Forrester Research wrote that banks should hold off on deploying chatbots for banking as the customer experiences they offer today are often either poor or uneven, and while I agree with him that experiences can be inconsistent, I don’t believe banks should hold off. In fact, doing so might deter prospective customers, especially millennials, who favour self-service and ease of communication in today’s ‘on-demand’ economy. Instead, banks should understand that not all chatbots are created equal.

Chat Mining

Most of the startups providing chatbot technology today are trying to work from a Q&A and automate it – and that’s where the customer experience can feel inconsistent. That method is simply not as effective as chat mining, which involves dissecting real chat transcripts between real customers and customer service agents. By using chat mining techniques, banks can unlock the nuances of human speech (which is always evolving), and understand customer preferences and behaviours that lead to critical insights.  This method also allows chatbots to achieve a level of service consistent with a company’s best agent.

Emotional Intelligence

Critics argue that chatbot bankers lack the emotional intelligence to handle more complex scenarios, such as a consumer weighing a difficult financial decision. That’s a false choice, however. Chatbots are better at some things and humans are better at others, and humans will be better at handling emotional situations for the foreseeable future. It’s all about balancing the two. For banks, and for any business, the goal should be to have a chatbot that performs as well as their best performing chat agent. That would ideally apply to empathy too.

There’s a difference between understanding emotions, however, and emulating them. Customers generally know that they are dealing with chatbots and attempts to humanise them can often end in failure. One of the debates that’s taking place is around what type of personality chatbots should have. While people may find Siri’s humour entertaining in their personal lives, they might be turned off when interacting with a business. There’s a concept in artificial intelligence called “the Uncanny Valley” -- the point where engagement falls off and they have a repellant factor – an uncanny, almost-human creepiness.

The chatbot’s tone is important. It needs to be intelligent and efficient, friendly and helpful, but it shouldn’t pretend to be a person. Some companies are over-investing in personalities because they don’t have enough humans available to handle all inquiries. Is that really good for anyone? It’s a bad idea to mislead people, and large enterprises should avoid doing anything that will mess up their brand.

Digital personal bankers are the future

Research has shown that four in 10 Americans haven’t visited a branch in the last six months, which also explains the 5% annual closing of branches since 2009.  As such, banks can no longer rely on their friendly personal bankers to sell new products and services to consumers when they visit the branch. The gap has grown so wide that banks are losing the revenue opportunities that came so naturally through these in-person interactions. So, what can banks do to recapture these opportunities?

There is a logical evolution taking place, from in-person branch personal banker to digital personal banker. A digital personal banker utilises technology such as natural language interfaces (virtual assistant or voice) self-service or chat to engage and guide consumers along their banking journeys. Unlike a branch personal banker, a digital personal banker is proactive. It can anticipate what a consumer might want next, and leverage the complete context of the customer relationship (spending/purchase history, etc.) to deliver a superior and a more personalised experience.

Top Tips

Banks considering using chatbots as personal bankers, should consider whether the chatbot can do the following:

• Handle basic questions – As consumers conduct research and compare products and services, they often use the same queries such as “best credit card to pay off debt,” “best savings account,” and “best mortgage options.” Handling these questions with a chatbot is far superior to forcing a consumer to navigate the site for answers.
• Quickly escalate to live chat – for complex questions, especially during customer acquisition (buying a new service, for example), it’s important that consumers have the same continuous experience. The ability to escalate to a live agent increases the ability to convert customer sales and deepen relationships.
• Incorporate natural language– Banking customers still use the phone for certain scenarios such as bill payments, application status checks to stop payments, etc. These customers often experience frustration due to confusing phone tree menus and long wait times to speak to an agent, even though their query is relatively routine and straightforward. Using a smart IVR that understands natural language can handle these lower-complexity requests means that customers can self-serve, and live agents can be reserved for higher value queries.
• Leverage smartphone capabilities – Since these devices are both auditory and visual, banks can offer a true multimodal experience where customers can use touch as well as speech to complete banking tasks. For example, if a customer chooses to accept an offer for a card upgrade over the phone, they can receive the new terms and conditions via text on their smartphone, thus lowering wait times and making use of the phone already in the palm of their hand.
• Explore emerging channels – From a banking standpoint, messaging applications and chatbots open many doors for easier and conversional digital banking interactions in self-service or chat conversations. For example, in the event of possible fraud, rather than email a potential fraudulent activity, banks can send a message to a customer’s Messenger app, alerting them in real-time and instantly display charges to be verified. In the event of a dispute, they can enable a resolution by seamlessly integrating with chat agents.
• Meet security requirements – at minimum, chat conversations should be encrypted. If a chatbot can’t do that, it’s not ready for prime time.

Investments in the above-mentioned technologies today can bridge this separation between banks and consumers and also and ease transition to the digital personal banker. Chatbot bankers may very well be a key part of the way banks grow their customer base and deepen customer relationships going forward.