AI shopping traffic in 2026: what Adobe Digital Insights reports
Adobe Digital Insights reported that AI-driven traffic to US retail sites grew 138% year over year in May 2026, which makes AI shopping traffic a clear retail signal for an online store in 2026. Adobe's blog, read on 19 October 2026, adds that AI traffic to US retail sites has surged 1,324% since October 2024. Percentage growth describes direction, not volume, so neither figure says how many shoppers any one store should expect. The same Adobe report says its analysis covers more than 1 trillion visits to US retail sites. All of these are US retail aggregates. They describe the sector as a whole, and they do not predict any single store's traffic or sales. An owner or ecommerce lead should treat them as a reason to look at their own data and decide whether AI-referred shoppers deserve a separate line in reporting.
Adobe puts the longer view in its blog: AI traffic to US retail sites has surged 1,324% since October 2024. That is a very large move over a short period, and it is a move in traffic, not in revenue.
The quality numbers matter as much as the growth numbers. In May 2026, AI-referred retail visitors converted at a rate 54% higher than non-AI traffic, while a year before that, AI visitors converted at nearly half the rate of non-AI visitors. Adobe's blog adds that AI-referred retail shoppers in May 2026 spent 53% longer per visit and browsed 23% more pages per visit.
Adobe says its analysis covers more than 1 trillion visits to US retail sites. That is a large base, and it is still a view of the US retail sector. Your store may sit above or below the average, and the numbers cannot tell you which. To see why assistants send shoppers to one store and not another, read how assistants choose which products and stores to recommend.
Counting AI shopping traffic in Google Analytics and Search Console
Google Analytics help, read on 17 October 2026, says the default channel group includes an AI Assistant channel for visits from sources like ChatGPT, Gemini, Deepseek, Copilot or Grok, so in 2026 that channel is the first place to look for AI shopping traffic in your own data. The same help page, read on 19 October 2026, says the channel excludes Google's AI Overviews and AI Mode, and that Organic Search includes them. Google Search Central, also read on 19 October 2026, says sites appearing in AI Overviews and AI Mode are reported in the Performance report in Search Console, under the Web search type. Two places therefore hold two different kinds of AI-related visits. A store that reads only the AI Assistant channel will miss visits that came through Google's AI features, and a store that reads only Organic Search will mix them with ordinary search visits.
Google's own help page is specific about the channel: the default channel group includes an AI Assistant channel for visits from sources like ChatGPT, Gemini, Deepseek, Copilot or Grok. It is also specific about the limit: the AI Assistant channel excludes Google's AI Overviews and AI Mode, and Organic Search includes them. For the Google features, Search Central says sites appearing in AI Overviews and AI Mode are reported in Search Console's Performance report, Web search type.
One more detail changes how you read the numbers. Google Analytics help says the session traffic-source dimensions in the Traffic acquisition report describe a user's most recent session. That is a narrower thing than the origin of a whole shopping journey, so a report built on those dimensions answers a session question, not a journey question.
Server-side requests are a different thing again. OpenAI's crawler documentation lists separate user agents for search indexing, model training and fetching a page live for a user. A fetch is a request to a page. It is not a referral visit and it is not an order. A user agent is a name a client declares about itself, so treat it as a declared label and never as a verified identity. A browser-like visit is likewise not proof of a person. For the wider gap between counting visits and tracing results, see AI traffic analytics versus AI attribution.
What to measure so AI shopping traffic is more than a headline
Adobe Digital Insights found that a year before May 2026, AI visitors converted at nearly half the rate of non-AI visitors, and that by May 2026 AI-referred retail visitors converted at a rate 54% higher, so AI shopping traffic can change character within a single year and deserves its own measurement in 2026. A store can keep four counts apart. Referral visits are shoppers who arrive from an AI assistant. Agent fetches are page requests made by software, sometimes to fetch a page live for a user. Attributed orders are orders your analytics credits to an AI source. Incremental lift is the extra revenue that would not have happened without AI. The first three can be counted from your own data. The fourth cannot be read from an attribution report, because an attributed order is not proof of extra revenue.
Keeping them apart protects you from two common mistakes. The first is adding agent fetches to referral visits, which inflates the count of shoppers. In agentic commerce, software may request pages for a person, and that request should never be merged with a visit that a shopper made. The second is reading an attributed order as extra revenue. Some of those orders may have happened anyway through another route, and only a comparison with a baseline can say how many.
Arbling's home page says Arbling separates AI-referred shoppers, shopping agents and crawlers, and shows which products they touched and which journeys became orders. It also says Arbling follows the visit from the AI referral to checkout and links it to a real order. A linked order tells you which journey ended in a sale. Whether that sale was additional is a separate question, and no single report settles it.
For where those journeys lose shoppers after the first click, read where AI shopping journeys lose revenue. If you want these counts for your own shop, get early access.
Where Arbling fits. Arbling separates AI-referred shoppers, shopping agents and crawlers, and shows which products they touched and which journeys became orders. It follows the visit from the AI referral to checkout and links it to a real order.