Macy’s Built an AI Shopping Assistant, Customers Using It Spend 4.75x More Per Session

In March 2026, Macy’s launched Ask Macy’s, a conversational AI shopping assistant powered by Google Gemini, deployed across all digital channels simultaneously. A/B tests conducted with roughly half of all site visitors revealed a result that executives announced publicly at Shoptalk 2026: customers who used the chatbot spent 4.75 times more per session than those who did not.

For readers who want to go deeper, Grace Certified offers a catalog of AI-evaluated business scenarios.

The context matters. Macy’s is a 168-year-old department store chain that had spent several years navigating a digital transition while facing structural pressure from e-commerce. The company’s chief customer and digital officer, Max Magni, articulated the strategic frame clearly: every AI investment had to “either solve a customer issue or unlock a customer opportunity or unlock better effectiveness.” That framing eliminated novelty-for-novelty’s-sake and focused development on measurable outcomes.

Ask Macy’s was the result of applying that standard to the discovery problem, the friction between a customer’s intent and the 800,000+ product catalog that Macy’s manages across apparel, home, beauty, and gifts.

The Overhaul: AI as Transformation, Not Feature

Ask Macy’s operates as a conversational layer across Macy’s website and mobile app, powered by Google Gemini. Customers describe what they’re looking for in natural language, an outfit for a summer wedding, a gift for a mother who likes cooking, a suit in navy under $300, and the assistant responds with curated, shoppable recommendations.

Alongside Ask Macy’s, the company launched Complete the Look, a virtual try-on feature enabling customers to upload a personal photo and receive tailored outfit suggestions. For customers using this feature, session spending was approximately five times higher than non-users, a figure disclosed by Magni in mid-2026.

The two features are not isolated experiments. They represent a deliberate architectural shift: instead of making search incrementally better, Macy’s rebuilt the discovery experience around intent-based AI interaction. The investment was in changing how customers engage with the catalog, not in adding another filter or recommendation widget.

The Results: Verified at Scale

The 4.75x spending figure came from A/B testing conducted on approximately half of Macy’s digital visitors, a sample large enough to be statistically meaningful for a retailer generating hundreds of millions of online sessions per year. Bloomberg reported the data on March 26, 2026, citing the disclosure directly from Macy’s leadership.

The mechanism behind the lift is visible in the product design. A customer browsing independently navigates a catalog optimized for search, functional, but passive. A customer interacting with Ask Macy’s enters a guided experience: the assistant narrows intent, surfaces relevant products, and extends the session with complementary suggestions. Each successful recommendation increases the probability of a subsequent one, compounding session value in a way that traditional browse-and-search paths do not.

For a retailer managing hundreds of thousands of SKUs across multiple categories, this is a structural shift in how the catalog generates revenue, from passive availability to active discovery facilitation.

What This Means for Traditional Retail

The Macy’s case carries a specific signal for organizations in retail, hospitality, and any other sector where the core challenge is helping customers navigate complexity to find what they want.

The 4.75x result was produced by an AI deployment that changed the customer’s relationship with the catalog, from search to conversation. The technology required (large language models with multimodal capability) is commercially available. The architectural decision required (rebuild discovery around intent, deploy across all channels simultaneously) was organizational and strategic, achievable by any retailer with the willingness to treat AI as a transformation rather than an addition.

Macy’s made that decision at 168 years old, in a sector under significant pressure, with a public commitment to measurable outcomes before deployment. The 4.75x is the measured result of that commitment. For board members evaluating digital transformation investments, it is also a benchmark: what AI-driven discovery can produce, in retail, with verified methodology and named organization disclosing the result.