Last week, four large retail chains, Gap, Dollar General, Ulta Beauty and Kohl's, spent part of their quarterly earnings calls talking about artificial intelligence. Not as a side project, but as a core part of how they plan to run stores, manage warehouses, and sell products over the next several years.
Ulta named a former Domino's executive, Kelly Garcia, as its new chief technology officer in July, tasked with pushing the beauty retailer's AI and technology work further. Kohl's said shoppers who use its new AI shopping assistant, launched with Google's Gemini technology ahead of back to school season, are already converting into buyers at a higher rate and spending more per visit. Dollar General, whose sales grew 5.2 percent this quarter to 11.3 billion dollars, said it is piloting AI systems meant to run across its whole business, not just one department.
Why now? Because every retail executive is under pressure to have an answer ready for their board. Aaron Cheris, who leads Bain & Company's retail practice, put it bluntly: everyone has to show investors what they are doing with AI.
That pressure does not always lead to good decisions. Research on "AI washing," where companies overstate how much AI they actually use, has drawn attention from regulators who worry that some of these claims are made for investors rather than customers.
Still, the numbers behind this shift are real. Bain's research across retailers suggests that AI powered personalization, things like smarter product recommendations and better search, can lift total revenue by 5 to 10 percent.
Separately, Adobe has been tracking how many shoppers arrive at retail websites after asking an AI tool like ChatGPT for advice. That traffic has been converting into sales at meaningfully higher rates than shoppers arriving through normal search engines or email links, meaning AI is not just cutting costs behind the scenes, it is starting to bring in paying customers directly.
This creates two separate jobs for retailers, and most are only doing one well. The first is customer facing: tools like Home Depot's Magic Apron, Walmart's Sparky, and Ulta's own shopping assistant, all designed to answer questions and recommend products the way a helpful store employee would.
The second is less visible but arguably more important: cleaning up product data, descriptions, and inventory information so that both AI shopping assistants and outside tools like ChatGPT can actually find and recommend the right products in the first place.
For any business selling physical products, the lesson is not to rush out a flashy AI chatbot. It is to make sure product information, inventory data, and customer records are accurate and well organized, because that is the foundation every other AI project depends on.
Retailers that skip this step end up with AI tools that sound confident but recommend the wrong size, the wrong price, or an item that is out of stock, which damages trust faster than having no AI tool at all.