The Maze: Alexa for Shopping is becoming a traffic allocator inside Amazon. Similarweb's U.S. desktop index for assistant-originated product-page visits rose from 100 in June 2025 to about 590 in May 2026—roughly +490%. The line briefly peaked near 730 during the holidays, then settled at almost six times its starting level. This is not total assistant usage, and the chart does not reveal absolute traffic. It captures something more commercially specific: an Alexa surface recommended a product, a shopper clicked, and an Amazon PDP gained a visit.
The assistant now controls a measurable path to the shelf. A product-page landing can start from search, a shopping guide, a comparison or mission-based discovery. It can also begin on another product's PDP: ask Alexa a question, receive a competing recommendation, click, and move. Traditional Amazon SEO tries to win the initial search result. Agentic discovery can re-rank the consideration set after the shopper has already entered the aisle.
The growth survived the holiday spike. The source line jumps sharply through November and peaks around 730 in December before falling back. Seasonality clearly helped. Yet May 2026 still sits near 590, not near the June 2025 baseline. That matters because it separates a holiday feature burst from a durable change in navigation. The intermediate monthly readings are approximate, but the reported endpoints still describe a 5.9-times expansion.
High-conversion users make the traffic strategically heavier. Amazon says Rufus users are more than 60% more likely to complete a purchase. That statistic is not a causal conversion lift; higher-intent shoppers may be more inclined to use the assistant. Still, it shows why Amazon is expanding the interface. The company says Rufus reached more than 300 million customers in 2025, and Alexa for Shopping now spans natural-language search, guides, comparisons, price history, deal finding and cart building.
Product data becomes the new merchandising surface. An assistant needs structured evidence: compatibility, dimensions, ingredients, delivery promises, price history, reviews and reasons one option fits a mission better than another. A weak PDP is no longer only a conversion problem after arrival. It can become a recommendation problem before arrival. Missing attributes give the model less to cite and a competitor more room to become the answer.
Amazon can monetize the recommendation layer twice. First, better guidance can reduce abandoned searches and move shoppers to a viable product. Second, Amazon Ads has begun describing sponsored prompts inside Alexa for Shopping. The resulting shelf mixes relevance, commerce data and paid influence inside a conversational interface. That raises the stakes for measurement: brands will want to separate organic agent referrals from sponsored exposure, and incremental switching from clicks that merely replace search navigation.
Why it matters: Amazon's familiar battlefield was keyword rank, retail-media placement and PDP conversion. Alexa for Shopping adds a fourth contest: recommendation eligibility. Brands should audit catalog completeness as machine-readable sales collateral, monitor assistant-originated PDP traffic separately and test whether the agent describes their products accurately. The winners will not be those that stuff the most attributes into a listing. They will be those that give the agent the clearest evidence for when their product is the right choice.
