Will the perfect chatbot continue to elude us?
The first chatbot was born in 1965 with a robot entitled ELIZA who could respond very basically in simple conversation. The technology was largely abandoned until its pointed resurgence in 2016, which can be attributed to major leaps in deep learning and natural language processing, as well as its relegation to messaging apps.
Presently, chatbots are able to speak rather fluidly. Their speech is much more nuanced than it was in the sixties, and they have a much better capacity to understand complex and colloquial language than they did only a decade ago.
Given these massive leaps, tech giants have sensed that the industry not only has the ability to generate revenue, but that it is also poised to grow exponentially. Facebook has already launched its Messenger Platform, which allows for companies to build their own chatbots for the social media behemoth. Google bought out a chatbot startup, API.ai, last year and continues to invest in the technology quite aggressively. What’s more, messaging apps, the main domain of chatbots, are quickly ascending into the hottest of mobile applications with over 2.5 billion people having at least one messaging app installed.
Facebook has been one of the biggest proponents of chatbots in customer service. During an F8 conference last year, Mark Zuckerberg showcased Facebook’s Messenger Platform with a quick and easy 1-800-Flowers demo, where he promptly ordered a bouquet of flowers without the hassle of opening up a webpage or calling in his order.
“[Messenger] is a simple platform. It’s powered by artificial intelligence, so you can build natural language services to communicate directly with people,” he explained before entering a few chat commands to complete the process. “I find it pretty ironic, because now to order from 1-800-Flowers, you never have to call 1-800-Flowers again,” he joked.
Human or digital?
Facebook’s investment in the technology seems to be prudent as the need for chatbots is on the rise. Aspect Software Research conducted an online survey in which one thousand US consumers, aged eighteen to sixty, answered a series of questions regarding automated customer experiences. Forty-four percent of the customers surveyed reported that they would prefer a chatbot to a human representative for customer service, if the process was feasible.
According to the same survey, an overwhelming majority, seventy percent, expressed that for simple interactions they would rather speak to a chatbot than a live person. Moreover, around forty-five percent of Millennials have already interacted with chatbots in some form, and are open to interacting with them more regularly. Despite these strong numbers, corporations are slow to adopt chatbots. According to a recent Intelligence Report released by L2, only two percent of index brands have adopted the use of chatbots on Facebook and only one percent have utilized the AI technology on Kik, a popular messaging app aimed at teens.
The incredible convenience of the 1-800-Flowers chatbot was not unduly recognized during Zuckerberg’s demo. Unfortunately, deep flaws and annoying bugs of the chatbot later surfaced. The technology sounded truly amazing, but in practice, it has yet to be perfected. Seventy percent of chatbots on Facebook fail, requiring human intervention at some point. This high failure rate may partially explain the slower adoption rate of customer service chatbots by major companies.
On the other hand, there have been customer support success stories that have helped propel the technology forward. KLM’s customer service bot, which allows for the purchasing of boarding passes, launched as a feature-rich, relatively bug-free bot, and was met with resounding success. The chatbot continues to evolve, becoming more and more human-like and personalized. In true chatbot fashion, KLM recently rolled out an emoji-led directions feature.
As companies continue to deploy chatbots, programmers continue to develop intriguing ways to circumvent common hiccups and users continue to test (and break) them, chatbot technology improves. We’re already seeing an explosion of chatbots, whether they be in the form of personal assistants, video gaming characters or language teachers. The more intricate applications, such as AI-powered customer service, are proliferating, albeit at a slightly slower clip. Most companies are open to deploying the cost-saving, headache-reducing technology.
Purportedly, eighty percent of US businesses want to implement it by 2020. Undoubtedly, companies and consumers alike are excited about the prospect of automating customer support interactions. The perfect customer service chatbot continues to elude us, however. As platforms and programmers innovate and tech companies invest, we can expect the rate of adoption to pick up speed. We’ll just have to wait for a few of the kinks to be worked out.