The Next Step for Retail AI Is Bringing It Into the Store

It’s time for AI to get physical.

At the National Retail Federation’s annual conference in New York this week, one opportunity discussed by both brands and start-ups was bringing AI analytics into stores, which remain the largest source of sales for many fashion businesses.

H&M and Tapestry each highlighted it as a top priority during a talk on the next technological game changer in retail. Tapestry is investigating a mix of computer vision and machine learning to gain insight on the demographics of customers coming into its stores, how they’re walking through the space and how they’re engaging with products and sales associates, according to Trang To, the company’s vice president of omni. H&M is similarly looking into what Ellen Svanström, its chief digital information officer, called “ambient intelligence” that makes use of sensors.

A key goal is to bring the same level of data analytics many companies already apply to e-commerce into physical retail. Online, brands are able to track what a consumer is clicking on, which products they’re searching for, how much time they spend on a page, where they drop off during checkout and more. Equivalent data on the in-store experience would have the power to transform store design, merchandising and the work of store associates, To said.

But high-quality data about consumer behaviour, or even the retailer’s own inventory, can be harder to collect and process in brick-and-mortar locations.

At NRF, a few businesses presented solutions, while H&M’s Svanström pointed to one low-cost technology that she said is also helping.

That would be “good old RFID,” she said. If you can collect all the data generated by the radio-frequency identification tags and “amplify that with an intelligent application layer,” you can truly start to provide insights, she noted.

H&M isn’t alone in seeing value in the concept. In 2023, American Eagle began implementing a new system in stores that combined computer vision, RFID and camera-outfitted sensors from the company Radar, which was also present at NRF, to give store employees a real-time picture of inventory.

Augmodo, a start-up featured in this year’s innovators showcase at NRF, makes sensors out of a store’s employees — or rather camera-equipped “smartbadges” they can wear on a lanyard. The company is able to analyse all the data employees capture as they move around the store, providing a real-time view that lets the retailer spot when inventory is out of stock or in the wrong location. At the moment, Augmodo is focused on fast-moving consumer goods, but the technology could be applied to any category, a representative said.

AI is also being turned to the area around a store. Arteli, another company in the innovators showcase, analyses local data ranging from consumer confidence and income levels to weather and style preferences to help retailers adapt their merchandising to the surrounding clientele. Chief executive Martin Gilliard, who previously held roles at Meta and Barney’s New York, pointed out that stores are where retailers often make their profit but can still be underutilised.

The same might be said of the wisdom of sales associates, a situation Tapestry is trying to correct with generative AI. The company has been using the technology to gather and analyse feedback from its sales staff, a process that normally happened through surveys and could require months to synthesise and implement, said Mandeep Bhatia, the company’s senior vice president of global digital product and omnichannel innovation, during another panel discussion on AI empowering retail employees.

Tapestry started with small trials but has since scaled the programme, which has resulted in some perhaps unexpected realisations.

“During the holidays … a lot of associates started using it. They got enthused and then they were like, ‘The music is really bad in stores,’” Bhatia said. “Based on that AI-generated summary of all the feedback, we saw that that was significant, and then we were able to change the music.”

Maybe more important than the musical choices was the finding that associate engagement has gone up 10 percent to 15 percent, according to Bhatia.

While it’s still early days for many companies exploring these applications of AI, it’s likely that companies will continue trying to find ways to use the analytical powers of AI on their most valuable selling spaces, and that the lines between digital and physical will continue to blur.