How brands, retailers are using AI — and what they won’t let it touch

AI Strategies // June 25, 2026 How brands, retailers are using AI — and what they won’t let it touch By Mitchell Parton Ivy Liu As retailers and brands continue to build out both customer-facing and internal functionalities using AI, the conversation has also shifted toward what executives don’t want it being used for.

This was the theme of many on-stage conversations with retail executives at the annual CommerceNext Growth Summit in New York City this week.

Executives are gaining a deeper understanding of where humans need to be in the loop, where they can completely take over certain processes, and where to continue investing in AI tools and the real returns they can deliver.

Kim Heidt, chief retail officer for Tecovas, said she and others at the Westernwear brand have been talking to store managers about how AI can serve as additional headcount for tasks such as managing email. But she said it won’t play a part in the interactions with customers in stores.

“I’m pretty adamant about no cell phones on the floor,” Heidt said, adding that because tech is already so embedded in everything, she is teaching employees to step away from it and learn how to have an authentic conversation with customers.

She said that over the past three or four years, as people have become more comfortable with technology, it has simultaneously become more challenging for store associates to walk up to customers and ask how their day is going. “As [employees] become farther away from those human connections, that’s something we really lean in on and train,” Heidt said.

“Not everyone is comfortable doing it. “ Other retailers are more comfortable bringing in AI as a tool in the stores. Last year, Ulta launched an AI tool that associates can use to get more information about products and trends to help customers.

Amiee Bayer-Thomas, chief retail officer at Ulta, said the company hopes AI technology will remove friction in the stores so employees spend more time with guests.

“It will be right there at your fingertips to help ease that interaction with the guests, but also to help you feel confident and armed with the information that you need in order to execute your job,” Bayer-Thomas said.

“If it’s not making our associates’ lives easier and allowing them to deliver on an elevated guest experience, and it’s not making our guests’ lives easier and doing the same, then we should be questioning and challenging why we’re doing it.”

On the creative front, Elizabeth Garry, vp of e-commerce at jewelry brand Pandora, said none of Pandora’s model imagery or on-skin imagery will ever be AI-driven. Still, Garry said Pandora is starting to test content where customers can see different views or 360-degree rotations of products, just again, while not being worn.

“We feel as though that’s the customer’s first interaction with being able to understand, connect and see themselves wearing the jewelry,” Garry said. “We believe that, psychologically, it’s really important for someone to look into the eyes of someone else and then see them wearing our pieces.”

Online, Garry said the company has employed an agent to better understand customers’ intents while shopping and guiding them through the process, and another agent that can deliver customer service while escalating to humans when the conversations aren’t going well.

“We are really trying to be intentional around how AI becomes a competency and not just a tool, and what the business objectives are that we’re trying to solve,” Garry said.

” No customer should live through a brand’s efficiency gains, and I think especially in a lot of the customer service realm, as an example, where there’s a lot of really proven use cases that can be very advantageous to the P&L, you have to be thoughtful around what that end-user experience looks like.”

Henry Spear, svp of digital and customer care for JD Finish Line, which owns Finish Line, JD Sports and Finish Line at Macy’s — which operates as a separate business — said the biggest thing that has changed for the company over the past six months has been the ability to take mass quantities of qualitative data from call centers and make that data digestible for its AI agents, feeding it back into its digital experience.

Spear said JD Finish Line is doing propensity modeling based on whether the company has heard from the customer before and trying to determine whether they should be assisted by an AI agent or taken directly to a human. In cases where their address has been difficult to deliver to, they may immediately be taken to a human when they reach out.

Similarly, Kristi Carlson, vp of digital technology for Kendra Scott, said the company now can do a lot more with its customer data from digital, stores and its loyalty program to customize t