‘If You Don’t Have an AI Strategy, You Are Gonna Die’

But there’s still a huge gap between the potential and the reality

Author: Andria Cheng

March 24, 2017

In the retail industry’s quest to please ever-more-demanding shoppers and cater to their individual needs to buy whenever, wherever and however they want, artificial intelligence is shaping up to be perhaps the most critical piece of technology merchants need to add to their stack.

“If you don’t have an AI strategy, you are gonna die,” said Devin Wenig, president and CEO of online marketplace eBay Inc., at the recent Shoptalk conference in Las Vegas. “The largest job function we’ve been hiring has been data analytics and engineering to extract pattern from the noise. The promise of personalization in our industry has been a letdown.”

The ultimate goal, he said: To create a unique shopping experience for every single shopper.

Wenig admitted there’s been a lot of buzz and skepticism surrounding AI but said “that doesn’t make it wrong,” adding AI technology has seen a real “breakthrough” in recent years. “I’d rather be early than late,” he said.

Last year, eBay bought three AI-related companies, including its October purchase of image-search firm Corrigon. Using Corrigon’s technology, eBay has introduced “Shop the Look” feature on some parts of its site including eBay Collective. The technology allows shoppers to hover over an image, say a chair in the picture of a living room, and the tool searches eBay listings to surface inventory that matches or is a close match to that chair.

Also using AI technology, eBay has been working on a personalized shopping assistant bot for Facebook Messenger.

“Instead of humans having to adapt their interaction, computers will adapt to humans in a way that’s natural for us.”

“The future is AI first,” said Jonathan Alferness, Google’s VP of product management, shopping and travel, at a separate Shoptalk presentation. “Instead of humans having to adapt their interaction, computers will adapt to humans in a way that’s natural for us.”

Google has been investing heavily the past 18 months in areas including machine learning, speech recognition and natural language processing, he said.

eBay and Google aren’t alone. No. 1 U.S. retailer Walmart said Monday it’s launching a tech incubator to invest in emerging technologies including machine learning and artificial intelligence.

Target’s first-ever startup accelerator program last year also included machine learning related startups.

The retail sector isn’t the only one hot on artificial intelligence. Healthcare, banking, auto and education are among other sectors talking about using machine learning to detect, for instance, fraud pattern in the financial sector.

In a November survey conducted for Infosys of 1,600 global IT and senior business decision makers from 10 sectors in US and six other countries, more than three quarters of respondents said AI is “fundamental to the success of their organization’s strategy.” More than three-fifths of the respondents in the study, released in January, said they believe their organization’s future growth is dependent on “large-scale AI adoption.”

The study also showed 84% of the respondents said they plan to train employees about the benefits and use of AI. By 2020, companies expect to see AI contributing a 39% average increase in revenue, the study said.

By 2035, AI is expected to boost businesses’ labor productivity by up to 40%, an Accenture study released in September showed.

However, while the promise of AI is there and retailers are looking at using the technology in areas including product search, recommendation, chatbots, and programmatic advertising, AI in many ways is still in the early stages and is full of glitches, industry experts said.

“The reason everyone gets so excited about AI is the science future vision,” said Tim Tuttle, CEO & Founder of conversational AI firm MindMeld, at a Shoptalk panel.

But, he warned, while there have been some amazing breakthroughs in conversational commerce and natural language development, most of the chatbots out in the market are still not good. Tuttle should know: He received his PhD from MIT's Artificial Intelligence Lab.

“I don’t think there’ll be millions of developers creating transformative bot experience because the level of investment that’s needed is too high,” he said, adding at the same time any company serious about this shouldn’t wait before it’s too late.

Late or not, AI, in the end, still has to compete with—and improve upon—the human touch.

Michelle Goad, CEO and founder of personal shopping app PS Dept., said in an interview that while the company, which partners with luxury retailers like Gucci and Fendi, uses AI for some functions, it’s human beings who do the job of actually recommending products to clients.

“Recommendation always has to be human,” she said. AI “doesn’t work.” 

See more coverage of Shoptalk:

Think Ultrafast Delivery is Only for Emergencies? Think Again

The Line Between Tech and Retail Blurs Even More

Retailers’ Top Worry in Trump Era? Border Adjustment Tax