How Machine Learning Is Changing Retail

With technology constantly evolving, artificial intelligence has quickly become a buzzword in making customer connections.

At natural-diamond brand Rare & Forever, artificial intelligence (AI) is now being used to grade stones and prescreen diamonds for imperfections.
“We’re thankful for the creation of the [existing] modern grading system,” says Chris Croteau, the company’s vice president of marketing and sales. “But it has its flaws.”

Rare & Forever’s proprietary AI grading — an alternative to traditional labs such as the Gemological Institute of America (GIA) — can see and do what people cannot, he maintains. “The cool thing about AI is machine learning,” the program’s ability to teach itself and make adjustments based on experience, he explains. “AI is 100 million times more accurate and 100 million times faster than the human brain can possibly comprehend.”

Using AI makes diamond grading feel objective, says Sheri Lien, who co-owns Lien’s Jewelry in Minot, North Dakota, with her husband Fred. Her store uses Rare & Forever’s grading system. “Consumers are so much more educated now,” she says. “At least, they think they are. It’s very difficult for a smaller hometown jeweler to communicate the differences between what people look at on the internet and the reality. The AI process really helps with that, because everything is graded by the same standards, all of the time.”

Better communication

AI also helps retailers engage with clients in other ways. Jeff Moriarty runs digital marketing for Moriarty’s Gem Art, his family’s store in Crown Point, Indiana, and serves as a freelance digital-marketing consultant for other jewelers. For his store’s email campaigns, he works with a company called Blueshift, using its AI to create effective subject lines.

“It comes up with what it thinks is going to be the best subject line to use for open rates,” elaborates Moriarty. “We tested it, and nine out of 10 times, their subject line resulted in higher open rates than our own.”

Moriarty also uses a chatbot. “Certain questions are answered automatically. Shipping, returns, and that kind of thing. Other, more specific questions are sent to me. I get a notification on my phone, and I start talking with them.” Advanced chatbots could be huge time-savers, he adds. “[With another] client, I added a chatbot, and it’s cut down on the amount of chat their customer service team has to do by about 75%. There’s still the option to talk to the operator easily, but there were, like, five questions that made up 50% or 60% of their chats. So it saves them a lot of time.”

Of course, not everyone feels a need for the latest technological advancements. Some of Lien’s customers simply aren’t worried about AI grading reports, for example. “They like to pick out what pleases their eye,” she says. “But people in their late 20s to mid-30s are much more into information. That’s usually when AI information comes into our sales presentation.”

Cost and benefit

For some retailers, the cost of implementing AI may be a barrier. “The starting package [for Rare & Forever’s grading system] is $50,000,” says Croteau, though he insists the benefits make up for the price. “With more accuracy, [customers] are actually able to get exactly what [they’re] paying for.”

The AI email service Moriarty uses costs him between $5,000 and $10,000 per month, but he, too, feels the conversion rates are worth it. “The higher the open rate, the more people will actually look at the email, which will drive more sales,” he explains. “It definitely works for us, because we sell higher-ticket items.” There are less expensive options out there, he adds. “MailChimp has a content optimizer, which is kind of a dumbed-down version of what we use. But it’s much more budget-friendly.”

Further down the road

AI tech will likely evolve based on the jewelry industry’s needs, says Croteau. “We only do rounds right now, but we could do fancy shapes. We could also shift it to [lab-grown] if we decided that we wanted to be in the lab business.”

Moriarty forecasts additional applications of AI technology: “Down the line, there’s going to be AI for product description or the content on the website. I could also see using AI to see how many times a piece of jewelry was looked at in a store. It’ll be exciting what happens in the next five or 10 years.”

Image: Shutterstock