Built Rufus to Dominate, Sparky to Catch Up, and Ask Macy's to Survive. All Three Are Winning.
"Every retailer is trying to figure it out one step at a time."
When Macy's launched Ask Macy's on March 23, 2026, it joined a race that Amazon and Walmart had already been running. Retailers experimenting with AI. The three assistants, Rufus, Sparky, and Ask Macy's now represent the retail industry's most concrete example of what AI-powered shopping tools can actually do to revenue.
Each has reported early performance numbers. And each number including a 4.75x spend lift, a 35% basket size increase, and $12 billion in incremental sales reveals something different about the retailer reporting it.
Rufus, Sparky, and Ask Macy's are all built to do the same thing. Convert browsing into buying. But the stakes behind each launch are not the same. Amazon deployed Rufus to deepen a marketplace that already commands hundreds of millions of shoppers.
Walmart deployed Sparky to close a decade-long e-commerce gap with Amazon. And Macy's deployed Ask Macy's to prove the department store still has a place in modern retail. Reading the numbers without that context produces a misleading comparison.
The urgency behind all three launches has a common source. As ChatGPT and tools like Perplexity began handling product discovery, retailers faced a specific risk: that shoppers would bypass their websites entirely and take their business elsewhere.
Macy's has spent the better part of a decade trying to reverse a sales decline. Net sales fell 2.4% last year. Comparable sales only recently turned positive, rising 1.5% for the full year, with Q4 delivering comparable sales growth of 1.8%, the third consecutive quarter of positive comparable sales growth the company had been working toward.
Ask Macy's, powered by Google's Gemini platform, launched across all of the company's digital platforms on March 23 after a beta launch in December 2025. During testing with about half of its website visitors over several weeks, Macy's found that customers who used the chatbot spent 4.75 times more than those who did not, the company reported to Bloomberg.
The most popular features are "complete the look," where the assistant suggests accessories to pair with an outfit, and a virtual try-on function that lets shoppers preview how items might look on them. Chief Stores Officer Barbie Cameron told Bloomberg that store staff have also started using the virtual try-on tool for customers who do not have time to try items on in person.
"Every retailer is trying to figure it out one step at a time," Max Magni, Macy's Chief Customer and Digital Officer said. "This is anybody's game. Nobody has cracked the code."
Ask Macy's was shaped by input from thousands of Macy's employees and was adjusted during testing to account for regional climate differences, brand preferences, and conversational tone.
Walmart launched Sparky in June 2025, positioning it as part of a broader shift from traditional keyword search to what the company calls intent-driven commerce. Sparky lives inside the Walmart app behind an "Ask Sparky" button and helps shoppers locate products, read review summaries, and plan for occasions.
On Walmart's Q4 FY26 earnings call, CEO John Furner disclosed that customers using Sparky had 35% higher average order values than those who did not. About half of Walmart's app users have now tried the assistant, according to Walmart US President David Guggina.
"Sparky is essentially helping us evolve from traditional search to intent-driven commerce," Guggina said. "From an economic standpoint, better discovery and higher conversion translates into bigger baskets and greater frequency."
CFO John David Rainey reinforced the framing on the same call, noting that Sparky users spend more while requiring fewer clicks to reach a purchase.
Amazon's Rufus is the oldest and largest of the three. They launched Rufus in beta in February 2024 and rolled it out to all US customers in July 2024 ahead of Prime Day. By the time Amazon reported Q4 2025 results in February 2026, the numbers had exceeded internal projections.
More than 300 million customers used Rufus in 2025. The assistant generated nearly $12 billion in incremental annualized sales, up from the $10 billion annualized pace Amazon had discussed after Q3. Customers who engage with
Rufus are 60% more likely to complete a purchase, according to Amazon's Q3 2025 quarterly report.
Amazon described Rufus's results as reflecting "an even stronger response than anticipated." The company uses a seven-day rolling attribution model to track purchases resulting from Rufus interactions, including delayed conversions.
What the Numbers Mean
The 4.75x spend lift, 35% basket size increase, and $12 billion in incremental sales figures are not directly comparable. They measure different things, use different methodologies, and come from retailers at very different scales.
Macy's figure comes from beta testing with self-selecting users, for a retailer still clawing back from years of declining sales, the goal at this stage is simply proving the tool moves the needle at all.
Walmart's is a basket-size lift, not a revenue total, a metric that matters most for a business whose e-commerce only turned profitable in Q1 FY2026. Amazon's is an absolute dollar figure built on 300 million users and a two-year head start. For them, success was never about proving viability, only scale.
Ali Furman, US Consumer Markets Industry Leader at PwC, told Bloomberg that about 40% of the top 20 US retailers by revenue have now deployed AI shopping assistants, with most launching in mid-2025 and into 2026.
"The effectiveness of these assistants varies," Furman said, "with the best models functioning as personal shopping agents that understand customer preferences and available product assortments, rather than merely serving as chatbots."
Beyond business context, the three assistants differ in how they are built. Rufus is trained on a custom large language model built primarily on Amazon's product catalog, customer reviews, and community Q&A posts.
It can now shop items in other online stores directly and complete purchases on behalf of customers through a Buy for Me feature, according to Amazon's Q4 2025 earnings materials.
Sparky and Ask Macy's are both powered by Google Gemini. Walmart layers its own retail-specific large language models on top of Gemini, training Sparky on item comparisons, personalisation, and shopping journey completion.
Macy's tuned Ask Macy's through input from thousands of employees, adjusting it for regional climate differences, brand preferences, and conversational tone.
What comes next for each assistant is already in motion. Amazon has expanded Rufus's Buy for Me feature to tens of millions of items across other online stores, allowing it to complete purchases on behalf of customers beyond Amazon's own marketplace.
Walmart CEO John Furner said on the Q4 earnings call that Sparky, currently available only in the US, will be expanded internationally. At Macy's, CFO Tom Edwards said the company has identified more than 35 AI use cases planned across the organization, spanning supply chain, merchandising, marketing, and customer-facing operations.
Each retailer started with a shopping assistant and is now building something larger.