Key Takeaways
- Amazon has begun selling the technology behind Alexa for Shopping as the "AWS Agentic Shopping Assistant," with Kate Spade signed up as its first customer
- It extends AWS's signature playbook of selling homegrown technology even to rivals into agentic commerce, with a 60-day launch path as the headline feature
- Amazon's message to "not hand the experience to an intermediary" forces every retailer to decide whether to own its AI shopping presence
Amazon Takes the Step of Selling Its AI Shopping Technology
Amazon said it's already signed up Kate Spade as a customer for its AI shopping technology.
www.cnbc.comOn May 27, 2026, Amazon announced that it would offer its AI shopping technology to other retailers through its cloud arm, AWS. It is a decision to open up the inner workings of the shopping assistant it has long refined on its own site to the broader retail industry, competitors included.
What it offers is a new service called the "AWS Agentic Shopping Assistant" (ASA). Amazon has packaged the architecture, starter code, and learnings developed for its conversational AI "Alexa for Shopping," and says retailers can launch their own AI shopping experiences tailored to their storefront, catalog, and brand "in as little as 60 days."
The notable detail is that the provider is AWS, not Amazon's retail arm. For the retail industry, Amazon is also one of the largest competitors. Offering the service through its cloud face is meant to ease the wariness retailers feel about sharing data with the giant.
Amazon's "Sell What You Build Internally" DNA, From Cloud to Logistics
This move is hardly a new idea for Amazon. Roughly two decades ago, AWS itself emerged from opening up systems built to solve internal infrastructure problems. The vast market of cloud computing originated as technology made for Amazon's own use.
The same pattern has repeated. From cashier-less checkout and warehouse operations to the more recent opening of its logistics network, carving out internally proven systems as "services" is Amazon's signature playbook. AI shopping has become its latest target.
The track record behind the offering is solid. Alexa for Shopping was consolidated and rebranded from its predecessor Rufus, and according to Amazon it drove nearly $12 billion in incremental sales last year alone and was used by more than 300 million people. The sequence holds: prove you can earn with it internally, then sell it outside.
First Customer Kate Spade and Its "AI Gift Concierge"
What this service concretely produces is shown by the case of Kate Spade, the early adopter. The Tapestry-owned luxury brand used ASA to build an "AI Gift Concierge."
The tool listens through conversation for the recipient, the type of occasion, and the shopper's intent, then recommends the right gift. It is designed to draw on real purchase-behavior data from Alexa for Shopping while preserving Kate Spade's own brand voice. Technically, it was reportedly built on the Anthropic Haiku 4.5 model on Amazon Bedrock.
The development timeline is concrete too. After about 2.5 months of testing, it went live in April 2026. Yang Lu, Tapestry's CIDO, expressed enthusiasm for the possibilities agentic commerce brings and noted the approach delivered the customization its customers needed.
The performance figure cited is not trivial either. According to AWS, conversational shopping sessions are expected to convert at 3.5 times the rate of traditional keyword search. Gift-buying, where the shopper has not yet decided what to purchase, is exactly the kind of moment where AI assistance that draws out intent through dialogue tends to shine.
What this case shows is that years of work building AI from scratch were compressed into about 2.5 months of testing. The brand supplies its own catalog, product knowledge, and tone, while borrowing Amazon's proven base for the underlying technology. That clear division of roles is precisely why even a luxury brand could go to production in a short window.
Why "Owning Your AI Shopping Presence" Is the Question
There is a clear philosophy embedded in Amazon's message this time. In the announcement, the company argued that retailers should build their own AI tools rather than hand the shopping experience to an "intermediary."
The basis it cited is the expertise each retailer holds. "Retailers already possess deep vertical knowledge about their products, customers, and categories that no general-purpose AI can match," Amazon said. The logic is that depending on a general-purpose "answer engine" dilutes the value specific to a brand and its customers.
A notable contradiction lurks here. Amazon itself runs a "Buy for Me" feature that pulls product information from other companies' sites to make purchases on a shopper's behalf, drawing backlash from some independent retailers whose products were listed without consent. Meanwhile, it walls off its own site from external agents and even won a court order against Perplexity. How far this asymmetric stance of "take from others, but protect our own" can underwrite the credibility of the offering remains an open question.
One Move Amid the Platformer Melee
The battle for AI shopping supremacy is already a crowded scramble. Place AWS's move in that context and its positioning comes into focus.
Google has built shopping features into its AI search results and is working with Shopify on an open commerce standard. OpenAI teams with Shopify and Walmart to surface products in ChatGPT, and Microsoft has added checkout features to Copilot. On the retail and marketplace side, Walmart, Target, and eBay are simultaneously building their own tools and riding along on outside AI platforms.
Sorting out how each company's approach differs makes Amazon's positioning clearer.
| Aspect | Building from scratch | AWS Agentic Shopping Assistant |
|---|---|---|
| Launch time | Years | About 60 days |
| Technical base | Designed from zero | Reuses Alexa for Shopping architecture and starter code |
| Data and catalog | Retailer-managed | Retailer-managed (kept by the retailer) |
| Brand voice | Fully custom | Customized to each retailer's brand |
| Provider | In-house / external vendor | AWS (positioned apart from rival Amazon retail) |
The market estimates back up how fierce this competition is. Accenture projects that more than 30% of online commerce will run through AI agents by 2030, representing about $3.1 trillion in transactions. If AWS can power the AI shopping infrastructure of retailers at large, it could become a new growth engine for the cloud business.
Challenges remain, however. Whether consumers are ready to hand off the task of completing a purchase to bots is still unclear. Some early movers like OpenAI have stumbled over technical bugs and the difficulty of onboarding retailers.
Conclusion
Amazon's launch of this offering marks an inflection point where agentic commerce shifts from "technology you build yourself" to "technology you buy and use." The 60-day launch speed may push retailers who have hesitated to build their own over the edge.
The question is where to place control of AI shopping. Hand it to a general-purpose engine, or own a shopping experience that leverages your brand and customer understanding? Now that platformers have started selling the technology, that question has become more concrete for every retailer.
