Macy’s AI Closes the Sale Before Shoppers Walk Away

Macy's AI

Seven in 10 online shopping carts are abandoned before purchase. Retailers have spent decades trying to close that gap through discount pop-ups, email reminders, loyalty incentives and recommendation engines. The rate hasn’t moved much.

    Get the Full Story

    Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.

    yesSubscribe to our daily newsletter, PYMNTS Today.

    By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions.

    Macy’s is trying something different. Ask Macy’s, a conversational artificial intelligence shopping assistant built on Google’s Gemini Enterprise for Customer Experience, works to resolve the uncertainty that drives abandonment before it happens, rather than chase shoppers after they’ve left. During beta testing, revenue per visit was 4.75 times higher among customers who used it versus those who didn’t, according to Google Cloud. Two months after launch, the assistant serves thousands of shoppers daily across Macys.com and the Macy’s app.

    That figure comes with a caveat. Ask Macy’s users skew toward intent-driven shoppers already close to a purchase. The 4.75x reflects revenue per visit among users, not a reduction in the overall abandonment rate.

    How It Works

    Ask Macy’s is multimodal, handling text and images across a catalog of more than 2.5 million SKUs. Rather than returning a grid of results, it asks follow-up questions. A shopper who says they need something for the office that transitions to an evening dinner gets asked about fit, fabric preference and occasion before seeing products. A shopper looking for a gift for a 10-year-old gets asked whether they want bright or muted colors.

    The virtual try-on feature lets a customer upload a photo and see how an item looks on them in different settings, including an office, a restaurant or outdoors. As shoppers share more, the recommendations narrow.

    “We wanted to show how technology can remove friction and elevate retail shopping for our customers, to help them feel guided, understood, and confident,” said Chad Westfall, Macy’s senior vice president of technology product development and customer experience. “We’re bringing the concept of hospitality to online customers at scale.”

    Advertisement: Scroll to Continue

    The Build

    Macy’s spent nearly six months on its own internal AI agent project before Gemini Enterprise for Customer Experience became available. The retailer scrapped that work and pivoted. “We realized we needed to pivot immediately,” Westfall said. “The pace of change and innovation in AI left no room to stand still, so we adapted. We frankly forwent some of our usual processes in order to go after this with all deliberate haste.”

    Dozens of engineers from both Macy’s and Google Cloud started daily standups in early February. Four weeks later, Ask Macy’s launched in beta to a small group of users and thousands of Macy’s colleagues. Within 24 hours it scaled to 50% of site traffic. A week after that, 100%.

    The Wider Picture

    PYMNTS Intelligence found in a survey of 5,841 consumers that nearly half used AI during their latest purchase journey, and 64% expect to use it more in the coming year. A separate PYMNTS Intelligence report found 43% of retailers are already piloting AI shopping agents, with the agentic commerce market projected to reach $1.7 trillion by 2030.

    PYMNTS reported Albertsons saw a 10% basket-size increase from its AI shopping assistant. Ulta deployed its own version on the same Google infrastructure alongside Macy’s.

    For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.