The rise of the chatbot (and do you need one?)

Chatbots are becoming ubiquitous on websites. You open the page and the chatbot appears in the lower right corner with a ding! to let you know it is ready and willing to help. 

According to Cornerstone Advisors’ “What’s Going On in Banking 2022” report, just 15 percent of community banks have already deployed chatbots with another 19 percent planning to invest and/or implement this year. The trend is obvious. The effectiveness, perhaps, is yet to be determined. We’ve all been on the receiving end of a useless chatbot, and, let’s face it, there are a lot of useless chatbots in the world. 

Don’t implement a chatbot simply because “everyone else is doing it.” Think about what you hope to accomplish and examine a chatbot’s features before making a decision.

Providing customer service

First and foremost: Chatbots are not a replacement for human interaction. They’re more like the triage, assessing a situation before someone more skilled comes in. If the chatbot can’t answer your customer’s questions (quickly), a human needs to be immediately available to take over.

And therein lies the problem for some community banks: Staffing a human to supplement the chatbot. If you can’t logistically figure out how this will work, then take a step back. You don’t want your customers to exhaust the chatbot’s resources only to find themselves “on hold” waiting for a human to enter the chat. 

To be effective at triage and reduce the burden on your customer service reps, chatbots need to have an in-depth library of answers to FAQs. If the chatbot can’t answer common questions, they’re not a time-saver and simply act as an extra step before talking to a human. 

The best chatbots are equipped with conversational AI. Conversational AI pulls the relevant words out of the chat to interpret the customer’s question. Some chatbots appear to have conversation capabilities, but a human (from your bank) has to write the scripted responses. AI-powered chatbots do not need a script and the responses to customer questions improve through machine learning. 

Collecting data about customer needs

A secondary use of chatbots is to collect data. Customers’ questions can provide a lot of insights. Questions like “Where can I find X?” might indicate the need to reorganize your website. Questions like “What are today’s mortgage rates?” are opportunities for a loan officer or marketing department. 

The data from a chatbot is readily available. Analyzing and knowing how to apply it is harder. Furthermore, chatbots would be only one piece of the puzzle. Understanding customer needs and using that knowledge to improve customer service or introduce a new initiative would also need to be consolidated with the data collected from phone calls or emails. 

Chatbots are only one part of a data management strategy. Customers will always turn to their preferred method of communication and that might not be a chatbot. Relying solely on chatbots to determine customer needs or plan new initiatives is relying on skewed data. 

Proactive chatbots

The most obvious use case for a chatbot is the pop-up (maybe with a friendly wave emoji) that asks “How can I help you?” and waits for the customer’s interaction. But chatbots can also supply information that the customer didn’t ask for.

For example, a chatbot within a customer’s online banking session might offer up a “Did you know?” chat; for example “Did you know that you can finance a remodel with a home improvement loan?” Unlike an email or website banner, the customer could immediately interact with the chatbot and ask questions.

Tips supplied by a chatbot should be targeted and personalized, just like any other marketing or messaging strategy. Bombarding the customer with information about every product and service offered will inevitably lead to an ignored chatbot. But supplying relevant tips at the right time can build trust and entice your customers to seek more information.

Much as chatbot vendors will try to persuade you that your community bank needs a chatbot, it might not be the case. There’s a fine line between helpful and annoying chatbots — and you’ve got to be convinced that your customers will benefit.