Last year, Facebook triggered a wave of hype surrounding chatbots and conversational interfaces in the wake of its rollout of 100,000 bots on Messenger. Sadly, they seemed to have bit off more than they could chew; the campaign was largely a disaster, with the AI behind Facebook’s chat platform failing to fulfill 70 percent of user requests. But the tech itself isn’t to blame — it’s Facebook’s premature and reckless roll-out. The tech is very sound, highly effective, and increasingly versatile. The industry doesn’t need any hype or overselling to prove that.
Chatbots have been around for a very long time and have progressively made their way into businesses because, perhaps more than anything else, they make practical and commercial sense, even when they’re not running on “real” AI.
A study released in May by Juniper Research reveals that businesses have saved $20 million in 2016 thanks to conversational interfaces, and the cost savings are on track to reach a whopping $8 billion by 2022. The crazy gap between now and then is indicative of how fast things are evolving and how sticky the technology is, immediately improving both business processes and customer experiences.
Chatbots are getting slick
Right now, the majority of the chatbots out there are primitive, doing little more than relaying your inquiry to a search engine and feeding it back to you. But the next generation of bots are popping up quickly, and it won’t be long until they’re a natural first-line intermediary between you and every brand or company, overflowing as a natural addition to our daily lives at home and work.
In fact, look to Asia and you’ll find that markets like China are already awash with chatbots, and we’re just catching up out in the West — typical when it comes to early adoption of emerging tech. Progress is being made on all fronts and in every direction. Last year, the Mitsuku chatbot won the Loebner prize and became the first of its kind to fool a panel of judges into believing they were chatting with a real, live human.
Whether you look at voice recognition, computer vision, natural language understanding, or any of the various branches that have a functional role in the evolving AI ecosystem, the advancements are steadily trickling out. For example, take the new machine learning approach dubbed deep cloning, which is specifically designed for making conversational AI a reality. Gartner projects that by 2020, which is less than 3 years from now, 85 percent of customer interactions will be managed without a human. What that means is that customer service professionals, for example, get to offload the majority of their mundane, routine, and menial tasks to their digital aides and virtual assistants, who can help structure and facilitate the kind of self-service care that customers are increasingly demanding.
It removes so many pain points in the process — imagine never having to be put on hold again!
Disruption can be positive
We should avoid taking Gartner’s projections to an extreme, however. Those in the customer service industry are not about to create mass job displacement in the wake of growing digital automation, because the real AI that would cause that kind of disruption just isn’t here yet. More importantly, we don’t know when it will arrive.
What we do know is that customer service professionals are in fact in love with the technology.
“I’ve worked in customer service for 20 years, and while the threat of getting replaced by chatbots is a concern, I am way more excited about using conversational user interfaces to extend my reach and abilities as a single point of contact at a startup to provide human insight at machine scale,” said Peter Swimm, a senior product specialist at bot startup Howdy and organizer of the Talkabot botmaker conference.
A survey released last month by Aspect indeed paints a picture where customer service agents aren’t depressed about the rise of chatbots. On the contrary, the technology is freeing up a considerable amount of their time, which is then channeled into better handling of the more complex questions and problems. The “easy question” management system has been relegated to their automated helpers.
“In the beginning, AI [artificial intelligence] can typically handle around 65 percent of users’ questions, so 35 percent go to the human customer service agent. But once we have more data, the AI becomes smarter. After one or two years, the robot can handle 75 percent to 80 percent of inquires. Of course, robots cannot fully replace humans at the moment. Even in two or three years, robots will still be assistants to humans,” Rita Chao, a Chinese chatbot creator, told eMarketer earlier this month.
Chatbots are enhancing things, not disrupting them.
And the next generation of chatbots are increasingly context-aware and proactive thanks to continual progress in predictive and sentiment analysis, which, along with natural language processing, creates artificial personalities that are becoming more and more human-friendly and helpful, which might foster better working relationship with their human colleagues.
Amir-Esmaeil Bozorgzadeh is the founder at Virtuleap, a sandbox for creative developers to showcase their VR concepts to the world, which recently ran the world’s biggest WebVR Hackathon. He is also the European partner at Edoramedia, a games publisher and digital agency with its headquarters in Dubai.