Amazon's shopping assistant passes 400 million customers as AI referrals grow

More than 400 million customers used Alexa for Shopping over the past year, Amazon said on October 29, 2026. U.S. customers who use the assistant spend 50% more per order than those who do not, CEO Andy Jassy said on the earnings call, according to PPC Land's account.

AI shopping is developing both inside retailers' stores and through assistants that send shoppers to other sites. Amazon's expansion shows the first route. A longer series from Adobe shows the growing role of AI referrals across U.S. retail.

For a merchant, the distinction matters. An assistant on Amazon may help a customer choose among listings before visiting a brand's own site. A referral from another assistant can bring that decision to the merchant's product page.

Amazon's expansion against a broader shift

Amazon reported 250 million Rufus users in October 2025, covering that year to date. Its February 5 results put usage above 300 million for the full 2025 calendar year.

On May 13, Amazon combined Rufus with Alexa+ to introduce Alexa for Shopping. By its July earnings call, it reported more than 350 million customers over the preceding 12 months. October's disclosure exceeded 400 million.

The Amazon timeline follows those disclosures and the product change. Each count retains its own reporting window, so the milestones show how Amazon's shopping assistant developed without treating them as a monthly usage series.

The Adobe line chart uses 23 monthly observations to track AI-driven visit share at U.S. retail sites. By August 2026, that share was 1,709% above its October 2024 baseline. The line rises through much of 2025, retreats from January to March 2026, then reaches a new high in August. Zero marks growth relative to the baseline, not an absence of AI visits.

Amazon's customer counts and Adobe's referral data describe separate activity. Together, they place Amazon's expansion in a market where AI plays a growing role in how shoppers reach products. Adobe's report supplies that market context; it does not measure Amazon's contribution to the referral trend.

Product information has to work in both places

For a brand selling on Amazon and its own store, an accurate answer needs to survive either route. A material description on an Amazon listing and the corresponding information on the brand's product page should support the same answer about what an item can do.

A shopper asking whether a jacket works for wet weather needs clear information about its fabric and limitations. That question remains the same whether Alexa helps choose the item inside Amazon or another assistant sends the shopper to the merchant.

Amazon's spending comparison also needs care. It compares assistant users with non-users, so it does not isolate how much extra spending the assistant caused.

Fanatics shows what better recommendations can change

Fanatics reported on October 26 that its new Shop the Look experience generated 10% higher engagement than the previous version in a test spanning roughly 30,000 MLB products.

The experience uses OpenAI models to assemble coordinated merchandise from available stock. Product attributes and team information guide the suggestions. Merchandising rules constrain the combinations, and the looks refresh daily.

That is a useful example for a retailer with a large catalog. A recommendation has to make sense as an outfit and remain available to buy. Better descriptions alone will not help if the suggested size has sold out or the accompanying item belongs to a rival team.

Fanatics measured engagement with the experience. Whether those interactions produce more completed orders remains a separate question.

More assistants will read marketplace listings

Meta's October 27 announcement names Shopee and Lazada among Muse's early connector partners in Singapore. Meta describes product research and price comparison, with connector features arriving progressively.

For merchants on those marketplaces, the listing may become the source of an answer before a shopper ever opens the store page. Specific material descriptions and sizing information give an assistant a basis for explaining why an item fits the shopper's request.

Common questions

Does Amazon's spending figure prove an AI sales lift?

It compares assistant users with non-users. Those groups may already differ in how often they shop, so the figure does not isolate the assistant's effect.

What did Fanatics improve?

Engagement rose 10% compared with its previous Shop the Look experience. Fanatics describes a test across roughly 30,000 MLB products, not a company-wide sales increase.

Are Muse's Singapore shopping connectors fully available?

Meta describes a progressive rollout. Availability depends on the connector and the task a shopper wants to complete.