For some time now, the number of bank branches in the United States has been on the decline. In fact, with the rate of bank closures doubling every 3 years, the last US bank branch could close by 2034. From HSBC’s 441 branches in the UK, 114 branches are being closed today, and since 2015, High Street banks and building societies in general have closed an astonishing 5,200 branches.
There are a number of reasons for this trend. The ever-growing popularity of online banking means that customers are increasingly doing their banking remotely. At the same time, the rise of mobile payment platforms and fintech challengers is eating into the profits of traditional banks. While during the pandemic consumers rapidly developed digital habits that they have maintained even as branches have reopened.
But, there’s another factor that’s often overlooked in discussions surrounding the demise of bank branches: the increasing popularity of chatbots and conversational AI. As the number of bank branches has fallen, the use of these conversational services has increased. In part due to COVID-19, users of Bank of America’s bot Erica increased by 67% in 2020. This growth has persisted and Erica now serves 1.5 million customers per day. Other large banks including HSBC, Scotiabank, and Chase have since followed suit with their own conversational AI and digital assistant offerings. This rapid growth in conversational services raises a key question. In light of these trends, could chatbots eventually make physical bank branches obsolete? HSBC cited that some of their branches serve fewer than 250 people a week.
While it’s true that chatbots don’t offer a particularly helpful service at the moment (they’re too inflexible because they struggle to understand natural language), it would be a mistake to write off conversational technology based on these limited experiences. In a recent survey, 40% of banking customers who chose to speak to a human did so because they couldn’t find the information they were looking for online. In a sense, bank branches continue to survive not because customers feel they offer an irreplaceable service, but because the digital alternatives are not yet advanced enough to cater to their needs. This is soon set to change. The next generation of conversational AI technology is rapidly advancing and will soon eclipse what chatbots are capable of.
| Chatbots | Next-generation conversational AI | |
| Natural language ability | – Access to an FAQ or knowledgebase – Structured around multiple-choice dialogues | – Capable of responding to requests phrased in natural language. |
| Conversational experience | – Single-turn interactions – Simple, limited conversational design | – Multi-turn interactions – Personalised conversations and responses |
| Ability to learn | – Reliant on reactive dialogue trees – Deductive processing | – Utilises machine learning and makes sophisticated use of context – Learns from conversations |
| Handover experience | – Customer forces to repeat themselves to human agent – Restricted when users can handover | – Agent provide with full conversation history and a summary – Routes customers at any time on demand |
| Banking tasks supported | – Handles basic requests | – Handles complex requests: e.g. changing customer details, fraud reports, sophisticated transaction querying |
| Channel Support | – Primarily text based – Limited support for voice interactions | – Text and voice – Accommodates requests made naturally via speech |
The limitations of chatbots
Chatbots have been widely adopted by banks as a way to interact with customers, and as of this year, 19% of mid-sized financial institutions have deployed chatbots. The technology is cheap, easy to deploy, and can resolve some simple banking-related queries. But, chatbots have their limits. They’re simply not as flexible as humans when it comes to responding to customer queries. And they also struggle to understand natural language and so fail to recognise what users are asking. This often leads to a lot of frustration on the part of customers.
Another big issue with chatbots is that they’re unable to grasp the context of a conversation. Without an understanding of context, you cannot get to the heart of a customer’s needs. And as a result, chatbots will often return irrelevant or unhelpful responses. When chatbots fall back rigidly on rules and keywords, conversations feel at best awkward. Above all, it means the bot is far less likely to be able to actually resolve the customer’s request. In the worst-case scenario, this can lead to a customer quickly reaching for the phone line, stepping into a branch, or giving up altogether.
Why are customers still using branches?
Examining why customers are still seeking the help of branch staff can help us to determine the direction that future conversational AI technology needs to take. For instance, 50% of customers who prefer to bank in a branch do so to access in-person advice. This would suggest that successful conversational AI technology needs to be capable of holding more in-depth and personalised conversations with customers. For the above-mentioned reasons, chatbots are simply not up to this task.
A smaller number of customers, 36.8%, prefer to bank in person because they feel it affords greater security. This is a harder anxiety to appease. As long as security standards are applied correctly, online banking is not in itself an unsafe activity. Rather, it is possible that alarming news stories about fraud, hacking, and cybercrime have made customers wary of managing their finances online.
Finally, people needing to deal in cash is a major factor in the continued survival of the bank branch. Cash transactions, payments, and interactions predominantly occur in branches. In fact, 53% of customers who use branch services said they did so because it affords easier access to cash. While there is less that conversational AI and chatbot developers can do to directly address this problem, it’s likely that the issue will decline over time as major economies transition into cashless societies. Again, the pandemic has done much to advance this trend. In 2021, only 19% of transactions in the US involved cash, and in the UK this number was 1 in 6. As cash transactions continue to decline, the role played by the physical bank branch is likely to decline considerably.
What will the next generation offer?
The next generation of conversational AI technology is specifically designed to address the failures of chatbots in order to provide a more natural and human-like experience. Unlike chatbots, this next-generation tech is able to understand natural language and respond in a way that’s more natural, engaging, and accessible for humans. This makes them much more flexible when it comes to responding to customer queries.
Next-gen conversational AI is also much better at understanding the context that surrounds a customer’s query. These services examine the full history and context of a conversation in order to select the most appropriate response. This allows them to provide more relevant and helpful responses. Building bots capable of holding more natural-sounding conversations is crucial in making digital banking appealing to those who still use bank branches to more easily access in-person advice. And a more convincingly human-like experience is more likely to win the trust and confidence of customers.
Likewise, banks need conversational AI tech that can offer a more personalised service and meet the unique needs of each individual customer. While one-size-fits-all solutions might eventually be able to automate a large number of customer service requests, a personalised service is crucial for winning over those customers who remain hesitant about switching to digital banking.
Fortunately, the next generation of conversational AI technology, such as the kind provided by action.ai, is designed from the ground up to address the failure of chatbots and to appeal to the kind of customers who would usually bank in branch. Our technology allows banks to provide a more natural and human-like experience that’s both flexible and helpful. It is exactly this new generation of conversational AI technology that is set to revolutionise the way banks interact with customers.
