- Successful businesses are based on positive relationships with their customers. How can the integration of technologies improve these relationships?
- Let's explore some reports of well-known sources about chatbots and how they influence businesses and customers.
- Do people like to do business with people or bots? What about the future?
Relationships matter to businesses. That’s especially true for lifestyle and consumer-driven brands. Therefore companies want to build positive relationships with their customers. The hope that technology can automate and streamline customer conversations was the foundation of the Chatbot hype. Although spruiked since 2016 by technologists, futurists, AI enthusiasts and early adopters, it seems that the Chatbot saga has already come to an end. Or maybe never really started. In this article, we have summarised the state of the chatbot application in 2018 and where it's heading.
‘Chatbots are the new apps,” said Microsoft’s CEO Satya Nadella in 2016.
The Hype Around Chatbots
Early 2016, the hype around chatbots couldn’t have been higher. Within the first 6 months of the Facebook developer tools launch, more than 30,000 Facebook messenger bots alone were created. The world of chatbots looked promising.
“…the 2016 bot paradigm shift is going to be far more disruptive and interesting than the last decade’s move from Web to mobile apps.”
Tech Crunch, early adopters, silicon valley – everyone published, pushed and conferenced on the topic. Now, only 2 years later, many Facebook messenger apps are orphaned. Reason enough to reconsider if that paradigm shift really even happened or if the area of chatbots has already come to an end.
Turns out chatbots aren’t a new technology. In fact, there have been interesting bot-to-bot conversations going on since the mid-1960s! (Check this bizarre transcript of AI psychotherapist bot Eliza and bot Parry who mimics a human with paranoid schizophrenia talking! https://tools.ietf.org/html/rfc439) And since the 1990s, the Loebner Prize drives AI scientists to create the most human-like bot. For two years in a row, chatterbot ‘Mitsuku’ won the Loebner Prize competition. Mitsuku pretends to be an 18-year old female chatbot from Leeds who can play games and do magic tricks at the user’s request. Millions of people, predominantly teens, have chatted with Mitsukoon Kik Messenger, Mousebreaker and Skype and conversed with the bot on relationship advice, bullying or other personal matters. Mitsuku had over 14 million interactions on Kik in just over 2 months.
As some chatbots like Mitsuko are so popular, it doesn’t surprise companies and brands wanted to jump on the bandwagon. “Any advertiser who doesn't sit bolt upright after reading that doesn't understand the dark art of manipulation on which their craft depends”, stated technology reporter Christopher Mimsin the Wall Street Journal in 2015. So the race of building bots began. A whole industry started to appear around bot development, especially messenger bots. Business Insider wrote,“Messaging apps are bigger than social networks,” noting that chat had surpassed social networking in monthly active users.
Figure 1 Chatbot Magazine, 2016
But apps are hard to program, plus people don’t download apps anymore (As of June 2017, 51% of US smartphone users downloaded zero apps per month; 75% downloaded two or fewer.) On the other hand, people already had the apps they needed and to replace them the bot needed to be better than what they already have. Replacing bots however, is difficult. For example, it is highly unlikely that someone would replace Google maps with a bot.
Other Pitfalls of Bots
Another huge problem of messenger bots in particular are difficulties with the processing of text. People feel that the bot just doesn’t get them. Bots understand words but not their meaning. In addition, conversations are not linear – they can be complex layers of a variety of topics. And AI for human language is difficult because languages are rich and dense. Computers have been invented to process data not to understand written or spoken word – in fact, this is one of the biggest challenges to date for computer scientists. AI has been researched since the 1970s but only now, in 2018, Google is launching their human-like AI voice system, Duplex. So the question is, why would someone replace existing customer support or FAQs that are working and accurate with something that is inaccurate and requires more effort to use?
The vision of customers chatting to brands through an automated chatbot was driving the chatbot hype for companies back in 2016. Jo Allison, consumer behavioural analyst at research firm Canvas8, which has published several reports on chatbots, found out that 90% of all social media messages to brands are usually ignored. The remaining 10% of messages are only responded to after 10 hours on average. Bots could help close the gap. But in reality, they turned out to be only a novelty. Besides the initial 6 months of hype, they never took off. People just gave up on them. Never revisited the downloaded messenger apps again. Ditched the conversations with AI. And why is that? Because the conversations with a bot happened at the wrong time, in the wrong place, and the wrong way.
Where from here?
Digit’s Ethan Bloch summed it up, quoted in a recent Inc article:
“I’m not even sure if we can say ‘chatbots are dead,’ because I don’t even know if they were ever alive.” After all, he said, no one can point to a chatbot that “all your friends were using.” Such a thing simply never existed.
The Hype is Over
The general sentiment: bots weren’t a thing after all. Pioneers in the field agree that it takes a lot of effort to get a bot up and running the right way and the best application right now is either as a disguised FAQ or as a new interface to process customer requests in a nice domain (e.g. “I have lost my credit card. Now what?”). The hope is, bots of the future will be an improvement over those we have now.The journey for the years to come is to let AI help people in providing better service. The better AI can process written and spoken word, the more data collected on the customer journey, the better bots may become. Organisations can then gradually move to an AI assisted human-led service.
For businesses, this means one less bandwagon to jump on. Learn from existing bots that survive the next year or so, meanwhile preparing onboarding and customer support for a digital roll-out. After all, people do business with people, not with bots – for now at least.
Figure 2 Chatbot Magazine, 2018