How AI can enhance the in-store customer experience

E-book excerpt: AI investments are key for e-commerce, but what about for in-store shopping?

David Weldon

February 16, 2022

5 Min Read

According to a recent retail shopping survey by Shekel Brainweigh, 75% of consumers say they would like to use autonomous stores, but only 2% say they have done so.

This is very typical in a new market, explained Guy Moshe, chief technology officer and co-founder of the firm. “It reflects that there is real promise there. People are excited to test it. So, this means that there is a real business opportunity behind it. And if it is implemented properly, then people like it.”

But there is such a thing as technology overkill in retail stores. It is important to use technology for specific customer benefits and avoid using it for its own sake.

“Don’t try to adjust the retail experience to the technology, but rather go the other way around. Just let the technology service your customers better,” stressed Rami Bar, head of business development at Shekel Brainweigh.

“With some retailers today, you will see very well engineered shops that are actually a technology demo rather than a retail center. They want to show customers that they have working technology. And nobody cares too much about the shopping experience,” Bar said.

“I believe that if a retailer goes this route, he will lose his customer and will see that the technology investment is not justified.”

Bar offered the example of a fully automated store, and things that management can do wrong.

“I need to flood it with light, which is unnatural, and not very pleasant. I need to retain certain SKUs and not others. And I need to make sure everything serves the technology best. This is not the right approach. You shouldn’t serve your customer based on the technology. It should just be a tool.”

That is the approach taken by Delek U.S. Holdings Inc., which has rolled out frictionless, artificial intelligence-powered self-checkout at its more than 70 retail convenience stores in Texas.

At the Delek stores, customers can walk into a store, select the items they want, place them in a kiosk tray, and have all items instantly recognized and simultaneously totaled in less than half a second, without the need to look for and scan barcodes.

Shoppers can then use credit cards, debit cards or mobile pay to complete transactions without touching anything but their purchase and the form of payment. Transactions can be completed in as few as 10 seconds, the firm noted.

Smart shelves

Another area where some retailers are using AI is with “smart shelves” – shelving with technology sensors that track what products are handled by customers, for how long, which are put back, where they are put back, and which ones apparently sell. Retailers investing in smart shelf technology are retrofitting existing shelves with four sensors located in the corners of the shelf. The shelf sensors are a 1.5mm thick metal plate, very easy to use, and easy to implement, explained Moshe.

“You can put up to 10 products on the same smart shelf. It recognizes each time that an item has been handled or removed, and it tells you which item has been taken,” Moshe said.

“You can track your inventory. You know what it is that you’re selling, so you can automatically replenish them. But you’re doing it at the shelf instead of at the point-of-sale. So, it is tracking interest in products, not just sales.”

“We are learning how shoppers are interacting with products. We learn what is the percentage of items being put back. Are they putting items back in place, or not in the same place? Basically, the heat map of the shelf,” Moshe said.

Retailers can also experiment with ideal product placement, and smart shelves can provide insights on what placement does to produce interest.

“What happens if I will take the Coke and put it in front of the Fanta, and not next to it? What will it do to customer interest?” Moshe asked.

Artificial intelligence can learn a great deal about a customer from product interactions, Dubois explained.

For example, “as people are picking up a product and turning it around to look at the ingredients, you now know that they are the type of person that cares about ingredients. And when they put a product back, that may mean there are ingredients in the product that they don’t care for,” Dubois said.

“All of these things allow me in the future to better target these people.”

AI and AR

In another example, Dubois said AI can tap augmented reality technology to support virtual mirrors that enable a customer to see how they might look in a particular clothing item if they were to purchase it.

“We’ve seen the magic mirror-type technique for a while, where someone stands in front of a mirror, they select the garment they want to be wearing, and the mirror uses AR to generate an image of that garment on the person. It lets them ‘virtually’ try it on,” Dubois said.

AI-based inventory tracking tools can also help a retailer better plan long-range product needs, as well as manage fast-moving consumer goods, such as food items. This can be especially helpful with any product shortages that might occur, such has been the case with a variety of items during the pandemic.

If a retailer could better understand what products are of the highest value to particular customer segments, it could better plan how it would respond to a shortage.

When a retailer has more knowledge on customers, their habits, and what they value most, it can more easily address supply chain issues that might arise.

To find out more about artificial intelligence in the retail sector, download our e-book – ‘AI in Retail: Reinventing customer relations

About the Authors

David Weldon

Freelance Reporter

Get the newsletter
From automation advancements to policy announcements, stay ahead of the curve with the bi-weekly AI Business newsletter.