Chatbots: All About Them & Why You Should Care

Live Customer Service can be daunting, regardless of which side of the interaction you’re on. When providing real-time customer support, there is significant amount of overhead and coordination required to be effective. If you're on the receiving end, you just want the provider to get to the point and help with your matter, whether positive or negative in nature.

Enter: Artificial Intelligence.
 
No, not the AI you remember from sci-fi films or incendiary news reports. This is AI along the lines of your friendly neighborhood Alexa and Siri—it lives online to field questions and search data to solve problems. This is the chatbot, and you may have already interacted with one without even realizing it.
 The term “chatbot” is used to describe any automated service that can communicate with real people, seamlessly, through a previously-existing and popular communication service. At present, the two most popular service-channel examples are Facebook Messenger and Slack; both of which offer API services that let chatbots communicate using the service. Facebook’s chat API allows for in-chat, guided product purchasing, and Slack ships out-of-the-box with its very own trainable Chatbot helper
 
Most of the recent "chatter" surrounding the use of chatbots has been in the retail industry. In a survey completed in December of last year by Retale,almost 60% of millennial respondents indicated that they had used a chatbot—and two-thirds of millennial respondents would actually buy an item through a chatbot without human interaction.

While retail may be the first to more wholly adopt chatbots, we think that there are great uses for chatbots beyond that industry. Banking/financial services, real estate, law, hospitality, medicine, and education are all fields that could benefit from the service provided by chatbots. Imagine searching WebMD's data through a chatbot, and having it proactively send you advice and health tips; or doing a preliminary search for a home using a bot, rather than waiting hours for email communication or searching on your own. 

Your industry, use case, capacity, messenger service, and business goals should all impact the decision on which type of bot to implement. There are an infinite number of ways to technically configure chatbots, including hybrids (powered by both real humans and an AI bot), so there is much to consider before diving in. Generally, however, bots fall into two distinct categories:

  • Scripted Bots are designed to guide a chat participant to one of several very specific outcomes, often with button-based prompts and very little ad-hoc input from the participant. Scripted bots are more difficult to scale, but offer the most simplistic interface for simple processes. These bots feel like a bot (think: automated phone systems) and should be limited to more black-and-white, yes-or-no workflows to be most effective.
  • AI-Backed Bots come equipped with the ability to process natural language with a service like Wit.ai, and can respond in kind with any number of “story” branches, leading participants to desired destinations, products, services, content, or purchases. These bots can be trained over time to accomplish any number of goals and, ultimately, feel closer to —if not indistinguishable from—a live support agent when properly trained.

While chatbots cannot exist on their own, they are intended to integrate with a large number of existing services and customer service solutions (basically any service that offers a chat-based polling API). They can even be configured to collect information from a participant and then hand it off to a live agent. Consider the overhead, time, and cost savings in simply being able to cut out the first 5 qualifying, repetitive questions about the participant in order to get them into the correct support funnel—and being able to do it in a way that’s nearly indiscernible to the participant, with no wait time, and all while expediting the experience for the participant in a convenient and efficient manner.

There are some great examples of chatbots in the wild already, and more popping up every day that could revolutionize the way we communicate with service-based products and companies. When properly poised, a chatbot can instantly reach an audience that’s consistently searching for easier ways to communicate with companies, and benefit both parties in the process.

So next time you’re chatting with someone in a real-time support setting, consider the fact that you may well be talking to a well-trained AI, and not "Sally Smithers" from Springfield. With the right application of technology and process, chatbots can replace a large amount of any repetitive real-time customer interaction, saving time and aggravation, and should very quickly become the predominant first-responder choice for live customer support in the coming years.