Though investment in the chatbot space continues to skyrocket, it’s safe to say we have entered a period of heightened skepticism. Gartner’s Hype Cycle for Natural Technologies, July 2021 (paywall) placed chatbots firmly at the nadir of the Trough of Disillusionment (“As an innovation does not live up to its over-inflated expectations, it rapidly becomes unfashionable and attention wanes.”). A recent study from GOMoxie found only 22% of consumers have a positive impression of chatbots. Some research firms have even gone so far as to call the current chatbot landscape a “failed revolution.”
None of this is a surprise if we place ourselves in our consumers’ shoes for a second. Over the past 12 months, we’ve all likely had the experience where a chatbot couldn’t answer our question. While we understand that chatbots aren’t all-knowing, it is still a frustrating and time-consuming experience. And as I’ve written before, today’s NOW customer expects a resolution to their problem as quickly and easily as possible.
Yet the most important part of the interaction — which can make or break our relationship not only with the bot, but with the brand — is what happens next. Did the bot devolve into a series of diverting next steps, like encouraging us to call into the general support line, as opposed to finding a solution for us in the channel we selected?
What Companies Get Right With Chatbots
A Userliike survey asked consumers to rate the most important things companies get right when using chatbots. Seventy-seven percent of respondents chose being given the option to escalate to a human agent the top answer. So the answer to ‘what makes a good chatbot’ is how well it can integrate with your human agents. But how do you set up that integration to optimize the consumer experience (and by optimize I do not mean “resolve the issue” as the chatbot might see it, i.e. deflect the customer to a different channel)? It comes down to clearly establishing what a bot should handle and how the handoff to a human agent happens when the bot falls short.
The first step is to determine the instances in which bot interactions are acceptable and perhaps even preferable for the customer. One such instance would be any inquiry in which a faster response can be facilitated by a bot. Speed to resolution is one of the key factors in customer satisfaction scores. Leveraging automation to accelerate your ability to answer customers’ questions should always be a priority. To this end, specific examples of inquiries that bots should be assigned include post-sale inquiries such as tracking, checking shipping statuses, fulfilling returns and pre-sale questions like identifying basic product information (e.g. "do you have this shirt in blue?").
Another instance of where bots can truly augment your overall customer experience is during employee off-hours. While most simple self-service bots may be little more than transactional, they can certainly be effective in the context of structuring your overall 24/7/365 customer service strategy, as long as you also provide the customer with information as to when they can get in touch with a human if need be.
A final instance is where the customer’s desired outcome is more general information. Inquiries such as “Where can I find more information about your company’s environmental footprint?” would be time-inefficient questions for your human agents to answer, especially if that information exists in full on your website. Understanding situations in which the customer would be most pleased by consuming FAQ-style content from bots results in both a better customer experience and less manual workload on your humans.
Related Article: A Good Chatbot Is Hard to Find
The Importance of Timing in Chatbot to Human Handoff
Of course these aren’t the only inquiries you receive from customers. According to Gartner, only 9% of customers can fully resolve their issues or accomplish their desired task via automation or self-service. This means that, for the other 91% of inquiries that automation can’t fully resolve, facilitating a superior customer experience will require a smooth transition to a human.
Clear parameters should be set up as to when the handoff occurs. Forrester recommends that, if automation cannot handle an inquiry after two attempts, then the customer should be turned over to a human. Additionally, any time a customer asks to speak to a human, their request should be fulfilled.
All relevant information should be provided to the human agent immediately upon receiving the inquiry. This includes pertinent information that was shared by the customer with the bot, the nature of the inquiry, and as much information about the customer and the product/service in question as possible. This step is a tricky balance to fulfill: The agent needs answers to any possible question that comes up in the chat, but they cannot be so inundated with information that they have eight tabs open over three browsers. It is this balance, in my opinion, that can make or break great customer service experiences. You can compound this issue by an exponential factor when you consider that most agents are interacting with three or four customers at a time on chat.
The final thing to keep in mind is that the agent receiving the escalated inquiry needs to act with extreme empathy for the customer. Chances are, if the bot has failed to answer their question, the customer is in an even more aggravated state than when they began the chat experience. Ensuring the agent can develop a personal connection and rapport with the customer can make the difference between developing an advocate who will speak phrases of your support team, and a detractor who may decide to do business with your competitors in the future.
Related Article: Chatbots Provide the Personal Touch. Here's Why That's a Danger
Meeting the Needs of the NOW Customer
To meet the needs of the NOW customer, customer experience requires a blend of automation and human interaction to facilitate both immediacy and successful resolutions in the customer's preferred channel. Ask yourself: Is your current chatbot helping or hindering in that regard? Make no mistake — I'm not suggesting the chatbot landscape is a “failed revolution.” In fact, I am incredibly excited to see what the chatbot space does to innovate and improve their offerings. At this moment in time, however, and for the foreseeable future, whether a chatbot is successful is largely defined by how it facilitates the handoff to the human agent. Brands who invest time and resources into solving this problem will be the ones who define the concept of “Incredible Customer Experience” in this NOW customer era.