In 2016 and 2017, chatbots were seen as one of the most important additions to the digital ecosystem, with some experts predicting that they would transform eCommerce, replace apps and save shopping malls, among myriad other capabilities. But lackluster artificial intelligence (AI) led most companies to put their plans on the backburner.
Also see: Can Chatbot Concierges Save The Mall?
Less than five years later, though, AI has advanced enough that some eCommerce executives are willing to give it another shot, looking at platforms such as Messenger and WhatsApp as the next step to reaching consumers.
“The messaging ecosystem is giant — lots of channels depending on where you are the world and what market you’re in — and eCommerce is shifting,” Matt Ramerman, president of Sinch for Marketing, told PYMNTS in a recent conversation. “Will dot-com stores go away? No chance, but the share of balance moving toward wanting to be able to shop a catalog with a smart … AI system, and being able to shop in messaging, is the future.”
Ramerman said he’s had conversations with several eCommerce directors at Fortune 500 companies that have all echoed this idea, which was surprising to him at first. “I was always thinking of it as a way to connect to … the catalog” by direct-messaging consumers with promotions and deals, he said. “But the fact that they’re thinking and seeing evidence that it’s the primary preference, the easiest way for consumers to shop, that’s a big deal.”
Walmart over the past month has made several moves toward expanding its conversational commerce capabilities. Last week, the box store giant acquired technology assets from conversation design startup Botmock, allowing designers, merchants, customer service and other non-technical teams to develop AI voice and chat experiences without the need for coding.
Last month, Walmart also revealed that it is currently testing the ability for customers to text Walmart their shopping lists.
Mike Myer, founder and CEO of communication platform Quiq, told PYMNTS in an interview that he expects conversational commerce to become a “significant portion” of purchases over the next five to 10 years.
“That’s what people do — they have conversations,” Myer said. “And the reason they interact with websites today is that they don’t have the opportunity to have a conversation. Given the choice between the website and a conversation, if the conversation was well-executed, I think people would be choosing that in a lot of cases.”
Until recently, Ramerman said, most AI conversations were limited to a set range of responses — which, as Karen Webster pointed out, made the shopping experience tedious, if not useless. But as Ramerman told PYMNTS, the adaption of complex neural networks to understand natural language and phraseology means that AI can understand up to 95% of consumers’ responses, even if the person has typos or a unique speaking style.
To be sure, he added, these networks are very complex and have to do a lot of quick decisioning to keep up with a conversation. But most companies also have a team of humans to handle requests that are too complicated or are misunderstood by AI, hopefully taking some frustration out of the process.
The easiest place to see this at work, Ramerman said, is in WhatsApp, where brands can have their own messaging channels. A customer might go to the L.L. Bean channel, for example, and say they’re looking for size 13 rainboots; the AI bot will then provide a selection of products to choose from.
“I think it’s pretty cool,” Ramerman said. “The website is kind of button-click static, but … there’s so much personality you can surface out of the bot, so we can bring the brand to life that way.”
PYMNTS research has found that 31% of U.S. consumers have voice-activated speakers in their homes, but Amazon currently dominates the voice AI market, with a reported 68% share.