Adobe says AI retail traffic surges as readability lags

Adobe reported a sharp rise in AI-driven traffic to U.S. retail websites and found that many retail webpages remain only partly readable by machines.

The analysis covered more than 1 trillion visits to U.S. retail sites, alongside a survey of more than 5,000 U.S. consumers about their use of AI in online shopping.

Traffic from AI sources to retail sites rose 393% year on year in the first quarter of 2026, the data showed. In March alone, AI traffic was up 269% from a year earlier, extending growth seen during the late-2025 holiday shopping period, when it increased 693% year on year.

The figures point to a shift in how shoppers find products online. Consumers are increasingly using large language models and AI chat services to search for items, compare offers and click through to retailer websites.

Adobe found that 39% of consumers in its survey had used AI for online shopping, and 85% of those users said it improved their experience.

Conversion data suggests traffic from AI tools is becoming more valuable to retailers. In March 2026, visits from AI sources converted 42% better than non-AI traffic, including channels such as paid search and email marketing.

That marks a reversal from March 2025, when AI traffic converted 38% worse than non-AI traffic. Adobe also found that 66% of respondents believed AI tools provide accurate results, a shift that appears to be feeding into purchasing behaviour.

Engagement trends

Visitors arriving from AI sources also showed stronger engagement once on a retail site. In March, engagement rates were 12% higher than for non-AI traffic.

Those users spent 48% longer on websites and viewed 13% more pages per visit. The pattern suggests shoppers coming through AI tools may arrive with clearer purchase intent than users from some traditional digital channels.

Visibility gap

Alongside the traffic and conversion data, Adobe published a benchmark on how readable U.S. retail websites are for large language models. Using its AI Content Visibility Checker, it assessed whether webpage content can be read by machines, with pages scored out of 100%.

An average homepage score of 75% means about a quarter of homepage content across the sector is not machine-readable. Category pages, such as men's apparel or appliances, scored 74%, while individual product pages scored 66%.

The lower score for product pages is significant because they often contain the information shoppers need to make buying decisions. Retailers typically hold thousands of stock keeping units, and weaker machine readability on product pages may reduce the chances of those items being surfaced in AI-led search results.

Other sections of retail sites showed mixed levels of AI readability. Store locator pages scored 73%, customer service or help centre pages 79%, contact us pages 81%, returns and exchanges pages 82%, loyalty or membership pages 78%, and FAQ pages 80%.

Adobe's data also highlighted a wide gap between stronger and weaker performers. While the average homepage score was 75%, the best-performing U.S. retail sites reached 82.5% and the lowest-performing came in at 54.2%.

That spread suggests some brands have moved faster than others to adapt their sites for AI systems. As AI tools become a more common route into eCommerce, differences in machine readability could affect which retailers and products consumers see first.

Consumer shift

The findings come as AI tools begin to act as an interface between shoppers and brands. Instead of starting their journey on a retailer's homepage or through a search engine, some consumers are now asking AI assistants to find a specific product, compare options or point them to the best price.

For retailers, that changes the value of webpages previously built mainly for human visitors and traditional search engines. If product descriptions, category information or support content are not legible to AI systems, those pages may become less visible as AI referrals rise quickly.

Adobe's data suggests retailers have made some progress, particularly on homepages and service pages, but product pages remain a weaker point. That may matter most because those pages sit closest to the transaction.

The survey and transaction data together suggest AI is moving beyond experimentation in online shopping. Traffic is growing, conversion is improving, and shoppers arriving through AI tools are spending longer on retail websites and browsing more pages.

Across key parts of retail websites, however, a notable share of content is still not machine-readable, with product pages averaging a score of 66%.