Wired recently published a piece by Arielle Pardes titled “What My Personal Chat Bot Is Teaching Me About AI’s Future”, in which Pardes relates what it’s like to use the app Replika. Replika is essentially a chat bot that learns to mimic and mirror your conversation the more you message with it, so that the user always has a “friend” to talk to.
“Originally, Eugenia Kuyda built Replika not as an AI to be your friend but one that would memorialize her friend, who had died in an accident in 2015,” writes Pardes. “The chatbot synthesized thousands of messaging conversations until eventually, it could reply in a way that sounded convincingly like Kuyda’s companion…. it provided a proof of concept: that the science-fiction idea of recreating a human life with artificial intelligence, à la Black Mirror, was possible.”
While Replika is more of a social experiment aimed at creating a sort of “digital clone” of somebody, the idea of AI-powered chatbots is one that’s already been adopted by marketers the world over — in fact, Intuit’s Firm of the Future blog listed bots as one of “6 Hot Trends Impacting Social Media Marketing”.
This is because AI chatbots allow for multi-million dollar businesses and small businesses alike to be present for their customers 24/7, and to help provide the always-on-call, conversational experience that users have come to expect (no matter the time of day). So how can you leverage artificial intelligence in your own social marketing campaign?
Invest in Conversational Experiences
Beerud Sheth, writing for MarTechSeries, makes his opinion clear by stating it his headline: “There Can Be No Denying: Bots Are Now Marketers”. He marks the importance of ‘conversational experiences’, calling it “merely the latest phase in the evolution of marketing”. He mentions that phase one and two were ‘outbound’ and ‘inbound’ marketing, respectively, but that this new phase involves “having one-on-one conversations with customers to ultimately build long-lasting relationships.”
“Outbound notifications will be transformed from the simple alert (that are an end in themselves) to the interactive messages that are the start of a new conversation!” writes Sheth. “Instead of receiving an alert notifying me of my package delivery, I prefer it includes an innate ability to instantly reschedule. Instead of an alert just confirming my flight ticket, I might like to view destination information at the same time.
No message will be standalone; each message will operate in conjunction with a rich, conversational experience. Even the frequency and timing of messages will be influenced by user preference and context. Nothing is ever the same.”
Sheth is right, and it’s because chatbots and AI are being designed to be “always-on” listeners. Not only do chatbots solve the “24/7 customer service” issue, but they are also, as Alex Debecker on the Ubisend blog puts it, “a fantastic tool to get insights on what your customers want and how they phrase the things they want.”
Marketers who currently use AI in their social marketing campaigns will generally be relying on chatbots — while those who don’t are all falling behind drastically in the marketing world. As time goes one, these chatbots will continue to grow more powerful, and the current downsides will diminish.
While AI are being used to fantastic degree in marketing, they’re not perfect yet. They still have a hard time understanding inference as well as emotion, which is extremely important in human to human job performance. The University of Maryland reports that Emotional Intelligence or (EI for short) is actually more important than standard IQ in the workplace, with EI being responsible for 58 percent of variations in professional and personal success factors, while IQ may only account for 4 to 25 percent of variances in job performance.
The reason that inference and EI are so important within the context of AI is that the machine needs to understand nuances in denotations and connotations of words and phrases that we take for granted on a daily bases. For example, “give me a break” can be a colloquial expression of exasperation as opposed to a request for a 15-minute break (different denotations associated with what is being said), while “thanks, you’ve been extremely helpful” can mean exactly what the words say they do, or the exact opposite of it if said in sarcastic manner (different connotations, derived from how something is said).
Because of these deficiencies, AI is currently unfit for deployment on its own, in most cases. A LiveWorld survey of over 200 marketers showed that in 2016, only 43 percent were using messaging apps for marketing purposes, while fewer than 50 percent felt that messaging apps, chatbots, and other existing two-way communication platforms enabled them to achieve their desired level of back-and-forth communication between marketer and customer.
The Answer: Combine Automation with Human Oversight
AdWeek ran an article by Laura Entis, titled “To Improve Chatbots, Brands Need to Combine Automation With Human Oversight”. She quotes Peter Friedman, the chief executive of LiveWorld, who recommends that brands use automated bots on platforms such as Facebook Messenger, but to do so with human handlers. She writes:
“This availability is what differentiates conversational marketing from digital marketing: it’s a shift from ‘always on’ to ‘always present,’ says Friedman. The Amazon Alexa speaker is a good example, in that it’s continually (but unobtrusively) on call, ready to respond and provide support. Companies should be “always present to support the conversation but only when the customer wants to have it,” Friedman says.”
Companies that want to embrace the always-on-call, conversational route that marketing is taking should begin looking seriously at AI. While it’s not perfect yet, it can be very effective with the right human oversight. Eventually these bots will do many of those jobs without human oversight — until then, we’ll have to lord over the bots.