AI shopping surpassed search on Prime Day, causing concerns for Amazon.

      Amazon established Prime Day to keep consumers engaged within its marketplace. This year, however, AI-driven shopping attracted top customers from external sources via chatbot recommendations.

      According to Adobe Analytics, US shoppers spent a record $26.4 billion across retail websites during the four-day Prime Day event. The more intriguing statistic is not the total amount but rather the origin of the most valuable buyers. For the first time, individuals who came to a retailer through an AI assistant were the most likely to complete purchases, as reported by GeekWire.

      Shoppers directed by AI chatbots were approximately 40 percent more likely to finalize their transactions compared to those arriving via search, email, or social media, Forbes observed. In contrast, just a year ago, this type of traffic represented the least effective converting channel. The rapid shift within a year is the key focus of the story.

      Analyzing the shift

      Adobe monitors billions of visits to US retail sites, and its Prime Day analysis revealed a significant disparity. On the first day, traffic from generative AI to retailers nearly doubled compared to the previous year, rising by 98.3 percent. This influx of visitors acted more like buyers than casual browsers.

      Shoppers referred by AI spent 49.9 percent longer on websites and viewed 20.5 percent more pages. They added items to their carts at a rate 33 percent higher than those coming from traditional sources. In essence, a channel that previously consumed a retailer’s time has now become its most effective.

      The larger pattern

      The data indicates that AI-referred traffic to US retail sites surged by 393 percent year on year in the first quarter. By March, this channel was converting about 42 percent better than non-AI traffic. In early 2025, AI visitors converted approximately 38 percent worse. This indicates a significant turning point.

      Impact on Amazon

      Amazon's entire business model relies on being the starting point for product searches, which is where its advertising revenue lies—one of the company's highest profit margins. The Adobe statistics suggest a shift in this starting point.

      If consumers increasingly begin their shopping journeys with AI assistants like ChatGPT, Gemini, or Perplexity and only visit a retailer for final purchases, the discovery process shifts up the line to AI companies, along with the associated revenue. Consequently, the assistant, rather than the marketplace, becomes the primary point of attraction.

      Amazon is actively responding by promoting its own Alexa+ assistant and Rufus shopping bot, and reevaluating its investment in the underlying AI models. However, the Adobe data suggests that general-purpose chatbots are already handling early-stage interactions, often before Amazon is involved.

      A change in entry points

      For the past two decades, the journey has typically started on Google. The same shift impacting AI shopping is also undermining Google’s dominance in search: users are opting to ask an assistant rather than enter a query. Marketers refer to this as GEO, presenting the notion that AI search is now the new SEO.

      The practical takeaway for retailers is clear: catalogs must be easy for machines to interpret, as chatbots can only recommend items they can understand. According to Adobe, many retail websites remain poorly optimized for machine readability, resulting in a disconnect between incoming traffic and the readiness of stores to accommodate it.

      There is a broader competition for commerce happening as well. OpenAI has already indicated its entry into the advertising sector, suggesting that a chatbot capable of both recommending a product and processing payments begins to resemble a marketplace itself.

      The situation in Europe

      A note of caution for readers outside the US: the Adobe data pertains specifically to American retail. Europe has not yet reached the same level of adoption of shopping assistants, and the EU’s Digital Markets Act is already altering how Google and Amazon rank and display products. If AI assistants do become the primary entry point in Europe, regulators who spent years challenging search consolidation will face similar challenges one tier higher. The entities controlling the assistants could take the place of traditional gatekeepers.

      This is a critical consideration for European retailers and brands. Visibility is transitioning from a search interface they have learned to manipulate to a framework that is opaque to them. The AI firms are currently defining the rules of this new space, largely based in the US and on their own terms.

      Considerations

      Two important cautions should be acknowledged. Firstly, the data originates from a single entity, Adobe, which provides both the analytics and AI tools profiting from these findings. Independent data would offer additional insights.

      Secondly, while AI referrals have grown significantly, they still represent a small fraction of overall retail traffic. A channel can be highly effective in conversion rates yet remain small, and a four-day sales event does not reflect a full year. The trend appears clear, but the scale is yet to be fully recognized.

      Nonetheless, the entry point into e-commerce is quietly evolving. For years, the focus has been on how to rank on Google or secure the Amazon buy box. The emerging question is more straightforward yet challenging: