AI shopping beat search on Prime Day, quietly threatening Amazon's discovery business model

AI Shopping Outperforms Traditional Channels for First Time

U.S. shoppers spent a record $26.4 billion across retail sites during the four-day Amazon Prime Day event, but the headline figure masks a more significant shift in e-commerce behavior 1

1

2

2

. For the first time, AI shopping assistants sent retailers their best-converting customers. Shoppers referred by AI chatbots were approximately 40% more likely to complete a purchase than those arriving from search, email, or social media 2

2

. Just a year earlier, the same AI-driven commerce traffic was the worst-converting channel available to retailers. That reversal inside 12 months signals a fundamental change in how consumers discover and buy products online.

Conversion Rates Tell a Dramatic Story

Adobe Analytics tracks billions of visits to U.S. retail sites, and its Prime Day data revealed lopsided growth in AI-assisted commerce 1

1

. On day one, generative AI traffic to retailers nearly doubled year-over-year, climbing 98.3% 1

1

. These AI-referred shoppers didn't just browse—they behaved like serious buyers. They spent 49.9% longer on site and viewed 20.5% more pages than visitors from traditional sources 1

1

. Most tellingly, they added items to their baskets at a 33% higher rate 1

1

. The pattern extends beyond Prime Day. Adobe Analytics data shows AI-referred traffic to U.S. retail sites surged 393% year-over-year in the first quarter 1

1

. By March, the channel was converting around 42% better than non-AI traffic 1

1

. Early in 2025, AI visitors converted roughly 38% worse 1

1

. The line has crossed decisively.

Amazon's Advertising Business Faces New Threat

The shift in product discovery threatens Amazon's core business model. The company built its empire on being the place where product searches begin, and its advertising business—now among its most profitable operations—depends entirely on that position 2

2

. J.P. Morgan expects Amazon's advertising to bring in approximately $83 billion in revenue this year and account for roughly a third of the company's operating income because margins run so high 2

2

. If shoppers increasingly begin with ChatGPT, Gemini, or Perplexity and only touch a retailer to check out, the discovery layer moves up the stack to the AI firms 1

1

. The assistant, not the marketplace, becomes the shop window. That's why Amazon has taken a different path than rivals like Walmart and Target, which have opened their catalogs to outside AI agents in e-commerce 2

2

.

Amazon Blocks External AI Chatbots Reshaping Online Shopping

Amazon has worked actively to keep AI shopping assistants away from its store, hoping shoppers will come directly to Amazon.com and use its own tools instead 2

2

. The company sued Perplexity over its browser that shopped on customers' behalf and won a preliminary injunction barring the tool from logged-in parts of its site 2

2

. Amazon has separately blocked ChatGPT's crawlers from reading its listings, even as it began buying ads inside ChatGPT to bring shoppers back 2

2

. According to J.P. Morgan data, agentic commerce drives less than 1% of traffic across every major online store, but Amazon's share sits lowest at about 0.4% 2

2

. That's by design, but it leaves Amazon fighting a two-front battle.

Alexa for Shopping and Rufus Counter External AI-Driven Retail Strategy

Amazon is pushing hard on its own AI-driven retail strategy. The tool launched as Rufus and folded in May into a service called Alexa for Shopping has drawn more than 250 million users, with monthly users up more than 115% over the past year 2

2

. Customers who use it while shopping are more than 60% more likely to buy, and Amazon Web Services said the tool drove nearly $12 billion in incremental sales last year 2

2

. CEO Andy Jassy said on the company's April earnings call that third-party AI agents in e-commerce weren't good enough yet, lacking a shopper's history and often getting prices wrong 2

2

. He argued people would gravitate to whichever assistant knew them best, stating "We are aiming to have it be the best shopping assistant anywhere" 2

2

.

Consumer Behavior Shifts Toward AI Agents in E-Commerce

More than half of U.S. shoppers are now willing to let AI handle the entire shopping process, including the final purchase, once their preferences are set, according to payments platform Adyen 3

3

. The shift extends beyond simple recommendations. AI agents in e-commerce now research products, compare options and complete transactions autonomously 4

4

. Zero-click searches have disrupted traditional discoverability and ranking tactics 4

4

. Online retail stores must be designed not for simple product searches but for detail-rich conversations 4

4

. Most consumers now consider AI-powered search answers their top digital source when making purchasing decisions 4

4

.

AWS Agentic Shopping Assistant Expands Amazon's Strategy

In May, Amazon Web Services announced the AWS Agentic Shopping Assistant, built on the same underlying technology as Alexa for Shopping and designed to let third-party retailers deploy comparable tools on their own sites 5

5

. The move reframes what Prime Day is actually testing. If Alexa for Shopping performs—if AI recommendations convert, price alerts trigger purchases and auto-buy runs without errors—Amazon will have a proof of concept it can sell to every retailer that runs on AWS 5

5

. The real stakes hinge on who controls the infrastructure layer of AI-assisted commerce once the event ends 5

5

.

What Retailers Should Watch

The big question long-term is whether Amazon can maintain its role as a primary destination for shoppers and avoid becoming just another selection on a chatbot's shelf 2

2

. For retailers, visibility is shifting from a search box they learned to game to a model they cannot see inside 1

1

. Catalogues need to be legible to machines, because AI chatbots reshaping online shopping can only recommend what they can read 1

1

. Adobe Analytics notes that most retail sites are still not very machine-readable, leaving a gap between the traffic coming and the stores ready to receive it 1

1

. The AI labs are writing the rules of that new shelf now, mostly in the U.S., and mostly on their own terms 1

1

.