Amazon’s new “Buy For Me” feature represents a stunning departure from the company’s decades-long strategy of keeping customers within its walled garden. By enabling AI agents to purchase products directly from brand websites when items aren’t available on Amazon, the retail giant is sacrificing short-term transaction revenue to maintain its position as consumers’ first stop in an increasingly AI-driven shopping landscape.
The feature, announced last week, is in beta for select U.S. customers, helps shoppers “discover and seamlessly purchase select products from other brands’ sites if those items are not currently sold in Amazon’s store,” according to Amazon’s announcement. What makes this particularly remarkable is how it contradicts Amazon’s traditional approach to e-commerce.
Breaking Its Own Rules
Scot Wingo, co-founder and CEO of ecommerce tech startup ReFiBuy and former CEO of ChannelAdvisor, called the initiative ‘bonkers,’ saying that Amazon has spent 31 years building a ‘fortress’ – a meticulously constructed a walled garden designed to keep shoppers within its ecosystem. Wingo points to the key elements of Amazon’s competitive retail moat, being the Prime membership program, Fulfillment by Amazon (FBA), its own payment processing, and countless other features specifically to ensure transactions flow through Amazon’s platform, generating fees and valuable data.
While “Buy For Me” keeps users firmly within the Amazon app interface throughout the entire process (as I explain with my own experience with the beta program below) it also abandons almost all of these sacred cows which are central to its core retail business.
The question is: why would Amazon seemingly kill its golden goose?
The AI Shopping Race
The answer likely lies in Amazon’s growing concern about AI shopping assistants potentially disrupting product discovery and purchase patterns – both of which are critical not only for its core business of selling physical products, but also for its ancillary but much more lucrative business of selling media space.
I recently wrote about in my Forbes article on Amazon’s patent for Alexa+, the company is racing to transform how consumers discover products through conversational AI, both within the web and app experience using its Rufus AI shopping assistant, and its nascent Alexa+ voice assistant.
Amazon says that Amazon Nova and Anthropic’s Claude models support the Amazon Shopping app’s agentic capabilities to complete the purchase from start to finish on a customer’s behalf.
In my analysis of Amazon’s ‘Nova Act’, I highlighted how Amazon is battling with OpenAI and others to control the future of AI-driven shopping agents, like Operator. I argue that the company that owns the AI assistant consumers trust to handle their shopping will dominate retail’s future.
“Buy For Me” represents another strategic move in this battle. By allowing its AI to complete purchases across the web, Amazon positions itself as the front door for all shopping journeys – even those that end on other websites.
My Experience With ‘Buy For Me’ As A Beta User
I was able to try the feature with footwear brand Rothy’s, which began selling on Amazon as a third-party seller just a year ago. After searching for “Rothy’s,” I found the Buy For Me widget a couple of lines down on the search results page.
What I discovered was particularly interesting: the variants in the Buy For Me widget appeared to be styles that aren’t sold on Amazon by the brand (The Casual Clog, Rothy’s Mens’ Clog, The Loafer Mule). This suggests the feature helps brands show their full assortment to Amazon customers without making all variants available there.
The experience of using Buy For Me was seamless, if not a little cold and joyless – I selected my size, clicked the button to buy, and received an order confirmation, all within the Amazon app without being taken elsewhere. The prices matched those on Rothy’s website. Once I placed an order, it appeared in a separate section from my regular Amazon orders.
In the workflow, Amazon also references a different feature that quietly appeared a few weeks ago: the ability to show customers products not available on Amazon.com when they search for a specific brand. Amazon said that the experience was designed to continue making shopping on Amazon convenient for customers.
“We’re testing bringing more selection and brands into our search results to help customers find even more of what they want and further improve our shopping experience for customers,” Rajiv Mehta, VP of Search and Conversational Shopping at Amazon said.
These two initiatives, when placed side-by-side, demonstrate that Amazon is prepared to take a hit on its core retail business in order to continue being the primary destination for consumers in their shopping journey.
Data Is The Real Prize With AI Shopping Agents
While Amazon seemingly sacrifices gross merchandise volume and merchant fees (at least while the program is in beta), the data collected through off-platform purchases provides the retail giant with increased visibility into consumer preferences. Every data point that Amazon collect on what a user’s interests, preferences, and behaviors are enables more sophisticated targeting options for advertisers and also informs Amazon’s own merchandising decisions.
Some industry experts commenting on LinkedIn believe this is primarily a data and advertising play. As Jason Goldberg, Chief Commerce Strategy Officer at Publicis noted, “The bigger share of a customer wallet they see, the better they can target ads. They also squeeze out other digital wallets. Get early signals and new products sales velocity, etc.”
This theory makes sense given Amazon’s growing emphasis on its advertising business, which has consistently outpaced its retail sales growth. By capturing data about purchases made on other websites, Amazon expands its ability to provide targeted advertising – potentially charging brands to influence which products are suggested through the Buy For Me feature.
Strategic Implications
For brands, Buy For Me creates an interesting opportunity. It potentially allows them to maintain direct customer relationships while leveraging Amazon’s massive traffic. Retailers using the feature can display their complete product assortment without providing their entire catalog on Amazon.
Amazon is clearly taking an “innovator’s dilemma” approach – focusing on aggregating shopper demand and delivering it, regardless of whether they own the inventory or collect the same merchant fees.
One thing is certain: in the battle for AI shopping dominance, Amazon has just made a bold, unexpected move that signals how seriously they’re taking the threat – and opportunity – of agentic AI in retail. The company appears willing to sacrifice some of its most cherished principles to ensure it remains the starting point for consumer shopping journeys, no matter where those journeys end.