We’ve been shopping online for almost three decades, but there’s still a massive downside to e-commerce for many customers: There are simply too many choices.

When you walk through the doors of a department store, you can go right to the section you’re interested in: women’s apparel, beauty, shoes, menswear, children’s clothing, and more. Once you get there, an associate can help you find your way to the products that most suit your needs.

By contrast, the almost infinite choice that e-commerce offers can sometimes be a detriment to decision-making. If you know what you want, a search engine can help you find it right away. But what if you don’t? You might need to enter multiple questions to find the right information, or enter a few terms and then start scrolling through search results … and then keep scrolling … and scrolling.

Despite years of ingenuity and massive investments into search technologies and product recommendations, many consumers simply give up. In fact, the average documented shopping cart abandonment rate today is still about 70%, so improving the customer experience to increase successful sales is a high priority for any online retailer.

Macy’s knows how to balance choice and service. A curated product assortment, visually organized store design, and personalized shopping experiences come together to define Macy’s as a modern marketplace.

Replicating that experience online was the motivation behind “Ask Macy’s,” a new digital shopping concierge and a major new leap forward in multimodal, agentic AI technology powered by Gemini Enterprise for Customer Experience.

With more than 2.5 million stock-keeping units (SKUs) in its product catalog, Macy’s knew it was crucial to solve the customer experience challenge while increasing sales and engagement.

“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, senior vice president of  technology product development and customer experience. “We’re bringing the concept of hospitality to online customers at scale.”

AI’s breathtaking speed
Agentic AI technology serves as an opportunity to upgrade the customer experience by providing customers with an easier and quicker shopping flow.

Crucial to the success of an agentic deployment is offering a level of service comparable to what customers get in-store, while providing a warm and conversational tone that is more inviting than a keyword search or just-the-facts shopping advisor. In other words, the bar for Macy’s was high.

At the same time, the sheer speed at which AI technology is advancing means that a project spanning months might be outdated by the time it launches.

After working for almost six months on its new website agent project, Macy’s learned about Gemini Enterprise for Customer Experience (CX). It had not been available at the start of its project, but Macy’s quickly recognized that it was ideally suited to the needs of Ask Macy’s.

“We realized we needed to pivot immediately. The pace of change and innovation in AI left no room to stand still, so we adapted,” said Westfall. “So, we frankly forwent some of our usual processes in order to go after this with all deliberate haste.”

One team, one dream 
Macy’s set a new target launch date for Ask Macy’s – the Shoptalk conference on March 24, 2026. At that time, that was less than six weeks away, so the Macy’s and Google Cloud teams sprang into action.

On February 9, the assembled teams – dozens of people from both Macy’s and Google Cloud – met for their first daily virtual standup meeting.

The teams collaborated on implementing Gemini

Enterprise for Customer Experience, integrating it with Macy’s existing systems, developing its interface, and deciding which features to include. The integrated team made decisions together to deliver a reliable, rich, feature-complete agent.

The team also incorporated feedback from Macy’s colleagues who tested early versions of the agent, helping make its responses warmer and more helpful. For instance, instead of answering a request for recommendations with: “Here’s a T-shirt for a 10-year-old,” Ask Macy’s now answers with something like: “10-year-olds can have so much fun with color – do you want a brighter or more muted color selection?”

Within just four weeks—an incredibly fast turnaround—Ask Macy’s launched in beta to a small percentage of the site’s users, plus thousands of Macy’s colleagues. Within a day, Macy’s scaled it up to 50% of site users, and then a week later to 100%. A little more than five weeks had passed since Macy’s had launched the joint project.

Would you like to try that on?

The agent goes far beyond a chatbot in that it’s not just text: It’s a multimodal agent, capable of handling text and images, and it also has a unique virtual try-on feature. If the customer uploads a photo, Ask Macy’s can show them how what they’re thinking of buying will look  on them – and it can even provide different backgrounds so they can see what it might look like at the office, in a restaurant, or trackside at the Kentucky Derby.

The Ask Macy’s experience has been live on the Macys.com website for desktop and mobile, and in the Macy’s app for iPhone and Android since late March. It provides customers with a seamless, concierge experience that can guide them conversationally through their shopping experience.

Macy’s is seeing people engage more conversationally as they start to realize the agent’s capabilities, with prompts like: “I’m a 50-year-old petite woman. High-quality material is very important to me – I love good cottons and natural fabrics. Can you suggest an outfit that works for the office but transitions to an evening dinner?”

Ask Macy’s will then respond with an overall recommendation, such as: “For your petite frame, look for tapered trousers or straight-leg cuts with a comfortable mid-rise. A silk blouse or crisp cotton shirt adds a touch of high-quality luxury that remains breathable throughout the day.” Then it suggests specific products in different categories.

As the customer interacts with it, providing more information, Ask Macy’s will respond with specific tips. For example, after learning the customer is going to a conference in Las Vegas, it recommended they dress in layers and wear flats instead of heels.

Ask Macy’s also incorporates data from customer feedback and the product catalog, and in the near future it will be able to warn you if products run large or small and recommend a more appropriate size to purchase. And it can recommend related items that go well with what you’ve got in your cart, if you’re interested.

A more unified whole
Just two months after launching, Ask Macy’s serves thousands of shoppers daily, and that’s growing as customers begin to understand its capabilities.

And Ask Macy’s is already driving strong engagement with customers. During beta testing, early data shows that revenue per visit was roughly 4.75 times higher among customers who used Ask Macy’s, highlighting its potential to drive revenue.

As it develops, Ask Macy’s will help bring together digital discovery and customer support, complementing the in‑store experience that helps define Macy’s omnichannel approach. “We want to meet our customers where they are and how they choose to shop,” Westfall said. “What you’re seeing here is just the start.”