Think about the best relationship you’ve had with a salesperson. The one who knows you go for bodycon silhouettes, love nothing more than a good deal and you’ve already purchased the blouse of the season in blush—and burgundy. Armed with this knowledge, they can pluck just the right look for you from the markdowns rack, making you very happy and closing the sale.
Head to that same shop’s e-commerce site expecting the same experience, and you’re almost guaranteed to be disappointed—at least for now.
With artificial intelligence, retailers promise to deliver the best of both worlds.
“When you think about [AI], think about your best sales associate or your best online merchandiser and putting those capabilities online,” said Andy Narayanan, GM of Intelligent Commerce, Sentient Technologies.
AI, and specifically machine learning, allows computers to crunch millions of data points almost instantaneously and provide better and better results with each interaction.
Add natural language processing into the mix and you don’t need to be an engineer to train your laptop or phone to anticipate your needs.
That’s the potential of AI—it’s human, only better.
These capabilities and their real-world applications are the reasons AI was on everyone’s lips at the 2017 ShopTalk retail conference in Las Vegas. The hope is that AI will turn shopping as we know it on its head.
“Brands are pushing product on customers and hoping that they’ll buy them,” Narayanan said. “How would it be if shoppers, engaged with the AI, pulled products that they wanted and pulled experiences that they wanted?”
“We believe that the future is AI first.”
Jonathan Alferness, VP of product management for shopping and travel at Google, was just one among many who said it’s not enough to incorporate AI, AI must lead business practices.
And indeed, AI is already ingrained and many Google features, including Google Translate, Photos and Inbox, which enable users to understand foreign languages, search an image library and create email shortcuts.
But to help consumers get the experiences and curated product selections they want from retailers, the company is continuing to evolve Google Assistant, which can currently accomplish tasks around your house and through your mobile device—all via voice commands.
Alferness explained that in the future, the keyboard will be extinct.
“Instead of having to adapt their interactions and communication paradigms to computers, now computers will adapt to humans so we can speak and interact in a way that’s natural to all of us,” he said.
But the full power of AI will only be realized once our computers are able to stay one step ahead of us, like automatically adding the peanut butter to your grocery delivery order when it sees you’ve added jelly. Or prioritizing the trendiest boots when you search for cold-weather gear over those that might be closer but not cuter because it knows that’s how you roll.
“Many people are leaving your app because they can’t find what they want.”
The irony of e-commerce is that as successful as it is, it could be more so if it could better mimic the experience of heading into a brick-and-mortar location.
“The biggest companies in the world understand that the basic principles of in-store shopping also apply online: Understanding customers, giving a good experience and helping them,” said Amir Konigsberg, co-founder and CEO of Twiggle.
Instead, right now what the 1.7 billion people who are currently shopping online get is a frustrating experience that pales in comparison to in-store service.
Additionally, Konigsberg said the vast majority of products offered online are in a virtual no man’s land, languishing undiscovered because traditional search algorithms are most useful for the most common queries only.
Twiggle has developed a platform that it says understands product data and uses it to return search results that dive deeper into retailers’ product catalogs.
Ultimately, this all about facilitating purchases, or, as Darin Archer, director of product management for Watson Commerce at IBM, said, “pushing the customer over the edge.” So far though e-tailers have only been able to muster a gentle nudge rather than a sharp shove.
“We’ve done that really poorly,” Archer acknowledged. “Recommendations suck. We put you in a cohort of everyone. But with AI we can focus on you and have a recommendation of one for you.”
With the power of Watson, IBM’s Jeopardy-winning platform, retailers can determine context and deliver an experience catered to you.
“The promise of personalization in our industry has kind of been a letdown.”
eBay is also working to foster this level of personalized service, according to Devin Wenig, the company’s president and CEO.
In fact, personalization was repeated like a mantra throughout ShopTalk, as though just uttering the word enough times would break the spell hanging over retail today.
Through personalization, consumers are supposed to receive, well, personal experiences. Instead of being presented with a site full of stuff, through which shoppers are meant to tease out the one thing they’re looking for, sites and apps will recognize you, remember your preferences and present an interface that’s customized for the unique list of parameters you’ll enjoy.
Unfortunately, for the most part, that’s still out of reach. But Wenig said just because it’s not here yet doesn’t mean the age of AI isn’t coming. “I think skepticism is fine but there’s almost no way that three years from now AI won’t deliver a more deeply personal experience, and I’m not going to be late,” he said, referring to eBay’s bullish stance. “I’d rather be early than late.”
Already AI is transforming how the mega marketplace interacts and serves users. For instance, the company is in beta mode with ShopBot, an automated personal shopper that operates within Facebook Messenger. It’s also continuously improving its image database, which he said is one of the world’s largest. With the latest technology, the computer can compare an image of a reference item to one that was just sold, determine which converts better, and select the better performer as the reference the next time.
“I think that if you don’t have an AI strategy,” Wenig said, “you’re going to die in the world that’s coming.”