June 16, 2026, (Inside AI) — U.S. consumers arriving at retail sites via large language models like Google Gemini or OpenAI's ChatGPT are spending more time browsing and generating higher revenue per visit, according to fresh data from Adobe Analytics. The findings signal a shift in how AI is reshaping online shopping behavior.
Shoppers referred from these AI tools produced 53% more revenue per visit than those coming from traditional sources such as search engines or social media, the May data reveals. This surge underscores the growing commercial power of AI-driven recommendations.
Why AI Referrals Are Outperforming Other Channels
The revenue lift points to deeper engagement. When a large language model suggests a product, the user often arrives with higher purchase intent. They have already received contextual, conversational guidance that mimics a personal shopping assistant.
Vivek Pandya, director of digital insights at Adobe, explained the mechanism behind the trend.
"Retailers whose products show up in LLM suggestions are able to drive more personalization to shoppers who leave the platforms to complete their purchases on the native websites."
This personalization loop starts within the AI chat and carries over, making the shopper feel understood before they even land on the product page.
The Hidden Infrastructure Battle
Not all retailers are positioned to benefit equally. The data highlights an urgent need for brands to invest in AI-readable webpages. Structured data, clear product metadata, and fast-loading, crawlable content help LLMs accurately retrieve and recommend items.
Without such optimization, a retailer's catalog may remain invisible to AI models that increasingly mediate consumer discovery. This creates a new digital divide between AI-optimized merchants and those relying solely on legacy SEO.
Adobe's findings arrive as e-commerce platforms scramble to integrate generative AI. Amazon, Shopify, and Walmart have all launched or expanded AI shopping assistants. Yet third-party LLMs like ChatGPT and Gemini are emerging as powerful gatekeepers outside any single marketplace.
Competing Views on AI's Retail Impact
Some analysts caution that the 53% revenue figure may reflect early adopter bias. Consumers using LLMs for shopping could be more tech-savvy and affluent, skewing results upward. As AI tools become mainstream, the premium might narrow.
Others warn of attribution challenges. If a shopper interacts with an LLM but later visits the site directly, the sale may not be credited to AI. Adobe's methodology tracks referral traffic, but the full influence of AI on purchase decisions could be even larger.
Privacy advocates also raise concerns. The deep personalization Pandya describes relies on data sharing between AI platforms and retailers. How that data flows and whether consumers fully understand it remains an open question.
What Retailers Should Do Now
The immediate takeaway for brands is to ensure their product information is accessible to AI crawlers. This means implementing schema markup, maintaining accurate inventory feeds, and optimizing for natural language queries rather than just keywords.
Longer term, retailers may need to forge direct partnerships with AI providers. Being a preferred source in a model's training or retrieval pipeline could become as valuable as ranking first on Google.
The Adobe data captures a moment where AI is moving from novelty to utility in shopping. As LLMs become embedded in daily life, the retailers that adapt fastest stand to capture the most value.
