Chatbots don’t always create better customer experiences

The financial services industry has embraced digital transformation as the way forward, yet digitization on its own isn’t exactly synonymous with progress. Retail banks that transitioned from in-branch interactions with customers to digital-only relationships aided by key technology such as AI chatbots have actually seen a marked decline in overall customer satisfaction ratings, particularly among Millennials, according to a J.D. Power report. The findings are especially disconcerting for big banks, which have generally outpaced regional and midsize banks in digital engagement.

Reports like this likely won’t slow the adoption of AI in banking, but they should give lenders pause. The reality is that AI still holds great promise as a driver of growth in the industry. Still, unless automated experiences are delivered thoughtfully, they can end up having the opposite effect.

This fall, McKinsey’s Global AI Survey of financial services leaders revealed that nearly 60 percent of institutions have implemented at least one AI capability. On the customer experience front, AI-powered virtual assistants (also known as chatbots) were used by 32 percent of respondents to enhance customer service capabilities.

Indeed, AI chatbots have the potential to help banks accomplish a number of critical objectives simultaneously. They can facilitate a significant reduction in operating costs by reducing the need for human customer service personnel while augmenting sales by offering products and services that maximize customer relationships. They also allow banks to provide superior customer service 24/7.

However, in the rush to reach the AI promised land, many organizations have provided less-than-stellar experiences that leave some customers longing for human interactions. I’m one of them.

After a recent attempt to change one of my personal bank accounts using my bank’s mobile application, I found myself looking for answers from the app’s help feature. The first line of help was a customer service chatbot. I spent several minutes trying to explain what I wanted to do before ultimately deciding that I’d need help from a human instead.

I’ve had nothing but great experiences interacting with my bank’s employees. In this instance, however, there was simply no obvious way for me to end the conversation with the AI chatbot to seek live support. I had déjà vu to phone calls with interactive voice response systems when I had to repeat “speak to a representative” ad nauseam. If chatbots become just another iteration of IVR, customers won’t buy into the benefits.

So what’s the solution? I’m not suggesting that banks give up on their attempts to improve customer service through chatbot AI, but there are three ways to improve these applications:

Focus on customer experience during design

Great customer experiences occur when organizations can anticipate and meet customer needs within a completely frictionless interaction. AI chatbots can play an important role there. For example, chatbots can quickly access key customer details while human customer service agents would have to sift through information to determine what’s relevant. That difference alone can speed up interactions for the customer’s benefit.

AI chatbots should reduce the work customers have to do. We often think of chatbot interactions as text-based, but there are opportunities to present customers with choices in the form of buttons or menus, rather than forcing them to type. Moreover, chatbots that can ask and understand compound queries — while harder to build — can shorten interactions and reduce the burden on customers.

Personalize your chatbot

An AI chatbot is a reflection of your brand and should provide a branded experience just like your website or mobile application. Your chatbot should fit naturally into the customers’ extended experience.

However, don’t pretend your chatbot is human. Even the most sophisticated AI can’t replicate an interpersonal interaction — and your customers don’t want to be tricked. Our UX testing of these systems has revealed that consumers are much more forgiving of technology shortcomings when they’re acknowledged upfront. Instead of trying to disguise your AI chatbot, celebrate it as a convenience-enhancing tool.

Plan for “failure”

You won’t be able to solve 100 percent of customer issues through AI customer service solutions, so plan accordingly. Make it easy for customers to transition away from the chat interface to seek help from a human representative and ensure the interaction history informs the subsequent engagement.

Prior to the COVID-19 pandemic, 52 percent of retail bank customers were branch-dependent, according to the J.D. Power report. With that in mind, banks shouldn’t seek to replace human interactions entirely. Rather, they should attempt to enhance upfront interactions and have a backup plan should those attempts fail.

Banks shouldn’t expect AI chatbots to be effective in all circumstances. By designing your application within a broader customer service offering, rather than as the only solution, you’ll have more leeway to continuously optimize without the risk of frustrating customers.

Kelly Franznick is the co-founder and chief innovation officer at Blink UX.