How Banks Can Nail the Handoff from Chatbots to Human Agents

bank chatbot

Manager of the company texting in the office

As a customer reports an issue to a bank’s chatbot, it becomes clear that resolving the problem is beyond the bot’s capabilities. The customer is transferred to a human agent—only to find themselves rehashing everything they have already discussed with the chatbot.

This is a common source of frustration in the digital banking experience. While it may seem like a shortcoming of the chatbot itself, the problem often has less to do with the technology than with how the interaction is managed.

As Lea Nonninger, Digital Banking Analyst at Javelin Strategy & Research, detailed in the Humans in the Loop: Chatbots Are More Effective When Customers Know Agents Are Standing By report, optimizing this experience requires introducing human intervention at the right moments.

To do so, banks must understand the limitations and strengths of their chatbot offerings—and configure them to ensure conversations continue seamlessly when a human agent is needed.

The Right Type of Query

Most leading U.S. banks now offer some form of chat interface, especially through their mobile banking apps. Chatbots have become table stakes, as consumers increasingly expect a responsive, digital channel to help them navigate their banking needs.

Compared with traditional self-service banking, chatbots can offer a more intuitive way to get answers and complete routine tasks. They excel at handling straightforward questions, such as those about fees or transaction limits, and helping customers perform basic account activities, including checking balances and reviewing transactions.

Chatbots can also assist with common service requests like reporting a lost card or understanding a declined transaction. Some can proactively notify customers about unusual spending or potential overdrafts.

Taking it a step further, more advanced bots can guide customers through new digital banking features and direct them to relevant products or services.

Perhaps their greatest advantage, however, is their availability. Customer can interact with their financial institution whenever they need assistance, whether they’re using a phone or computer. Even when a question seems routine to the bank, getting a timely answer can be important to the customer.

Despite these capabilities, chatbots are not equipped to handle every customer inquiry. The disconnect between their capabilities and customer expectations becomes particularly apparent when questions involve more complex issues, such as financial health or loan options.

When a conversation reaches this point, banks need a clear path to human support.

“It’s important to make sure that the chatbot is still connected to human support bankers as a fallback option that is available for customers when they need it—especially when it comes to more sensitive topics and things that require a lot of complex details and context,” Nonninger said. “It can become difficult to convey complicated topics to a chatbot, especially via a small chat window.”

Keeping the Conversation Going

Even when a chatbot can’t resolve an issue, it can still play a valuable role in the customer service process. Ideally, it can gather relevant information, authenticate the customer, and route the conversation to the appropriate agent—without requiring the customer to start over.

“Getting the handoff between chatbot and human support right is important,” Nonninger said. “It’s essential to avoid repetition from the customer side, and making sure that bankers know everything that’s been discussed with the chatbot before to avoid customers having to provide the same details again or having to authenticate themselves again—when all of that has already previously been done.”

“It’s on banks to make sure that that doesn’t happen and that the humans are ready to pick up where the chatbot left off,” she said.

One common source of frustration in customer service is being redirected between teams or agents, with each new interaction requiring customers to repeat information they have already provided. Every additional handoff creates another opportunity for the customer to disengage before the issue is resolved.

This is one of the key risks banks must manage when integrating chatbots into their customer service operations. Poorly coordinated interactions can undermine the convenience chatbots are meant to provide and, in some cases, damage the customer relationships.

“It’s important to not have customers be met with chatbots saying, ‘I’m sorry, I’m not trained on this yet,’ or just a generic error message,” Nonninger said. “Banks need to make sure that if a chatbot doesn’t understand or can’t provide an answer—and it will—that there’s some kind of safety net for customers so that they don’t have to abandon their chatbot conversation to seek out an alternative channel themselves.”

“If the chatbot does fall short, that experience shouldn’t end then and there,” she said. “They should then be guided to the next channel, whether it’s live chat or providing a phone number to call—just to keep that conversation going and not end with a frustrated customer that tried to seek help and wasn’t able to find it.”

Considering All Channels

The transition from chatbot to human agent also raises an important question—which channel should the customer use next? As digital service options have expanded, customers have developed different preferences for how they want to interact with their bank.

For example, when a chatbot can’t resolve an issue, the customer may be directed to a live chat agent. In many cases, that is an appropriate next step, but it’s not necessarily the right solution for everyone or every situation.

“It’s important to make sure that other channels are considered,” Nonninger said. “Maybe someone prefers to then pick up the phone after. It’s bringing that agency to customers, providing them with different options like, ‘Here’s a number you can call for your question,’ or ‘We can direct to live chat,’ or ‘You can schedule a branch appointment.’”

“I’m guessing in most cases where the chatbot fails to understand a query, customers would fall back into the live chat option initially as the natural progression, but it’s all about giving them options and agency to pick a channel that best fits their preference and circumstance.”

Exit mobile version