The Maze: AI shopping is no longer just a product demo hiding in a keynote. In Sensor Tower's 2026 AI work, GenAI sources send roughly 1.6-1.7% of web referral traffic to Walmart and Target in the US general-shopping category by May 2026. Still tiny. But that is the wrong comfort blanket. The more useful signal is the slope: both retailers moved from near-zero visibility in late 2024 to the top of the set in less than two years.
Walmart and Target are turning tiny AI referrals into visible demand signals. The Sensor Tower exhibit shows both retailers crossing the 1% line in late 2025, then separating from Amazon, Costco, and Temu by spring 2026. Walmart moves from roughly 0.1% in September 2024 to about 1.6% in May 2026. Target lands slightly higher by the endpoint. This is not a mass channel yet. It is a fast-forming lane.
Amazon's lower external share may be a measurement story, not a weakness story. Amazon's AI shopping assistant, now renamed Alexa for Shopping, was built to answer product questions, compare options, and recommend items inside Amazon's own experience. That can keep AI-assisted shopping in a closed loop. Sensor Tower's referral metric catches traffic arriving from GenAI sources. It may miss demand that starts and ends inside Amazon.
The middle of the set shows how uneven AI discovery still is. Costco finishes near 0.7% after a volatile spring 2026. Amazon sits near 0.5%. Temu stays below 0.3%. That spread matters because AI visibility is not a generic internet tide. It appears to reward brands and catalogs that answer engines can understand, trust, and recommend when shoppers ask broader questions.
Referral share is a weak proxy for commercial value. A LinkedIn commenter made the sharpest caveat: AI referral traffic is the visible outcome after a system discovers, evaluates, trusts, and recommends a retailer. Another noted that those sessions may arrive closer to purchase than ordinary search visits. The source does not prove that. But it sets the right operator question: not only "how much traffic?" but "what happened before the click?"
Why it matters: Retailers used to optimize for search rankings, paid placements, and marketplace shelves. AI adds a recommendation layer before the visit. The boring work gets strategic: clean product feeds, answer-ready content, review quality, trustworthy brand signals, and measurement that separates ordinary referrals from AI-assisted intent. Today it is 2% or less. That is exactly when channels are cheap to understand and expensive to ignore.
