When Mark Zuckerberg announced new tools letting developers build chat bots for Messenger, he demonstrated how a consumer could use a brand’s bot to order flowers: “You never have to call 1-800-FLOWERS again.” But behind the scenes, a human observed each interaction, ready to address any request the bot could not process.
As retailers embrace bots as the next big thing, humans still play an integral role in these interactions. The 1-800-FLOWERS bot integrates messaging platform LivePerson, which automates simpler tasks – ordering roses for Mother’s Day, for example – but depends on live agents to carry out more complex ones, such as when your flowers fail to arrive or you are overcharged.
“When consumers try to use a bot or virtual agent, there is a high error rate, which drives low customer satisfaction and ultimately a very expensive phone call,” LivePerson founder Robert LoCascio said. “There’s a lot of startup thinking around bots, but relationships are built with connections with a human.”
Yet dozens of brands are trying to build those connections with bots. Bank of America, Tim Hortons, and Business Insider are among several companies planning to launch bots on Facebook Messenger, seeking to engage the platform’s 900 million users, while H&M and Sephora have launched similar offerings on messaging app Kik. H&M bots provide fashion advice and mobile commerce, while Sephora’s offer beauty tips and tutorials.
Those investing in the bot space are optimistic about its future.
“Every brand that you can think of will have to have a bot to engage with their users, to enable them to communicate with them and find them,” said Beerud Sheth, CEO of bot developer Gupshup. “It is going to happen very quickly. I expect a bot explosion by the end of this year.”
The data suggests that consumers might at least be interested in the concept of bot-based retail. More than a billion people interacted with a chatbot in 2015, and nearly 40% of consumers would rather use messaging apps like Facebook Messenger for customer service than call over the phone. Asian consumers already interact with brands on messaging platforms: WeChat in China, KakaoTalk in South Korea, and LINE in Japan.
But there are also risks and challenges. Microsoft’s recent Tay debacle – in which the company failed to prevent its AI bot from tweeting offensive statements – reveals one potential pitfall. On a practical level, LoCascio also pointed out that bot developers for brands face a twofold challenge. They need a steady stream of data on how people interact with the brand and how they ask questions, but that transaction data might be private. And even if it is accessible, the developers might not be able to connect to backend systems that would let their bots offer helpful, account-specific support.
The appeal of bots lies in the fact that they offer brands a chance to be where consumers are: messaging. While smartphone owners use only a handful of apps, limiting the audience for branded offerings, messaging apps have become some of the fastest-growing social platforms. The number of global messaging app users is expected to increase from 1.4 billion in 2015 to 1.6 billion this year, and brands want to join the conversation.
“We see mobile messaging as a strategy for connection,” LoCascio said. “Ultimately, messaging is a framework for a greater relationship with the consumer.”