TL;DR
- Adobe Digital Insights published the report on 29 October 2026, based on over 1 trillion visits to US retail sites, 100 million SKUs and 18 product categories.
- In September 2026 AI-referred traffic grew 71% year over year, converted 90% better than non-AI traffic (it was 9% worse in August 2025) and added to cart at a 53% higher rate.
- Adobe forecasts AI traffic up 159% on Thanksgiving, 95% on Black Friday and 88% on Cyber Monday, and says the revenue-per-visit edge will double this holiday season.
- Adobe's own Citation Readability Score averaged 63% for US retail homepages; Electronics scored 79% and General Merchandise homepages 40%.
- The data is US-only and Adobe gives no AI share of total visits, so the 83% premium cannot be read as a share of revenue.
Shoppers who reached US retail sites by clicking a link in an AI assistant generated 83% more revenue per visit than other visitors in September 2026, according to a report Adobe Digital Insights published on 29 October 2026. The same AI-referred visits converted 90% better than non-AI traffic, the highest gap Adobe says it has recorded, and Adobe forecasts that the revenue-per-visit advantage will double over the 2026 holiday season. The figures come from Adobe's analysis of more than 1 trillion visits to US retail sites, so they describe the US market and compare traffic groups rather than measuring how much traffic AI sends in total.
What Adobe reported on 29 October 2026
The report, written by Vivek Pandya and published on the Adobe business blog, is built on direct online transactions. Adobe says the data covers over 1 trillion visits to US retail sites, 100 million SKUs and 18 product categories. A companion survey of more than 5,000 US respondents, conducted in September 2026, is mentioned as extra context on how consumers use AI, though the post does not publish results from that survey.
Adobe defines AI traffic here as users clicking on a link from a large language model. In other words, it counts visits that arrive from an AI assistant's answer, not every shopper who used AI at some point. Non-AI traffic is the comparison group, which Adobe describes as sources such as email, paid search and social.
The growth numbers
In September 2026, Adobe says AI traffic to US retail sites grew 71% year over year. Adobe began tracking this traffic in October 2024 and reports that it has grown substantially every month since then. Adobe also notes that it first saw an uptick in AI usage during the 2024 holiday shopping season.
For the 2026 holidays, Adobe projects AI traffic to rise 159% year over year on Thanksgiving, 95% on Black Friday and 88% on Cyber Monday. These are forecasts, not measurements, and they are growth rates against the same days in 2025. Because the starting base is not given in this post, a large percentage does not tell you how many visits that represents.
Conversion, engagement and revenue per visit
The second set of figures covers how AI-referred visitors behave once they arrive. Adobe compares them with non-AI visitors in September 2026:
- AI traffic converted 90% better, meaning visits from AI were more likely to become purchases. In August 2025, by contrast, AI traffic converted 9% worse.
- AI visitors spent 73% more time on site.
- They viewed 25% more pages per visit.
- They added products to their carts at a 53% higher rate.
- Revenue per visit was 83% higher, which Adobe calls the largest gap it has observed this year.
Adobe sums up the shift by saying that within 12 months, an AI visit has gone from being worth less than half of a typical visit to nearly twice as much. It forecasts that the revenue-per-visit advantage will double during the holiday season.
The table below collects the September 2026 figures Adobe published for US retail sites. Traffic growth is year over year; the other rows compare AI-referred traffic with non-AI traffic.
| Metric (US retail, September 2026) | AI-referred traffic |
|---|---|
| Traffic growth, year over year | +71% |
| Conversion rate versus non-AI traffic | 90% better (was 9% worse in August 2025) |
| Time on site versus non-AI traffic | +73% |
| Add-to-cart rate versus non-AI traffic | +53% |
| Revenue per visit versus non-AI traffic | +83% |
Citation Readability Score: how much of a page AI can use
The second half of Adobe's report shifts from traffic to content. Adobe uses a measure it calls the Citation Readability Score, which it describes as how much of a web page's content can be effectively understood and cited by AI. This is Adobe's own metric, and the post does not explain how it is calculated.
In September 2026, the average US retail homepage scored 63% on this measure. Adobe reads that as more than a third of homepage content being invisible to the AI systems and agents that send shoppers to retailers.
Results by category and page type
Adobe reports that Electronics leads all retail categories with an overall score of 79%. Within Electronics, category pages scored 79% and customer service pages scored 78%, which Adobe attributes to setup guides, troubleshooting articles and detailed specifications. Grocery followed at 55%. General Merchandise homepages scored 40%, and Adobe says their navigation-heavy, promotional layouts give AI less context to work with.
Adobe's conclusion is that the content AI reads best is content that explains. It names product details, buying guides, FAQs and return policies as the kinds of pages that AI can find, understand and recommend.
What the report does not tell you
Several limits apply, and they matter if you plan to use these figures in a budget discussion.
It is US data only
Every number in the report describes US retail sites. Adobe gives no figures for Thailand, Southeast Asia or any other market, so the size of the AI advantage in another country is unknown from this source.
The gaps are relative, and the share of total traffic is missing
The report compares AI visits with non-AI visits. It does not say what share of all retail visits came from AI in September 2026. A channel can convert far better than average while still being a small part of total traffic, so an 83% revenue-per-visit premium says nothing on its own about how much total revenue AI referrals produce.
Selection effects are likely
Adobe's own framing suggests why AI visitors perform well: by the time shoppers click through from an AI assistant, many have already done their research and are ready to buy. That means part of the gap may reflect who these visitors are and where they are in the buying process, not something about the channel itself. The report does not separate those effects.
The readability score is a vendor metric
Citation Readability Score is defined and measured by Adobe. It is useful as a direction of travel, but without a published method it cannot be compared with other AI-visibility tools, and a 63% score should not be read as an industry standard.
What retailers can check now
None of the following requires Adobe's tools. They are practical steps that follow from the patterns Adobe describes.
- Separate AI referrals in your analytics. Grouping those sources into their own channel lets you compare conversion rate and revenue per session with your other channels using your own data instead of a US average.
- Audit the pages Adobe says AI reads well. Look at your product detail pages, buying guides, FAQs and return policy. Check whether the main facts (price, specifications, compatibility, delivery, returns) are written in plain text on the page rather than locked in images or scripts.
- Look at your homepage honestly. Adobe found that General Merchandise homepages, with navigation-heavy, promotional layouts, scored 40%. A short block of descriptive text explaining what you sell and to whom gives an AI system something to work with.
- Measure before and after peak periods. If you change pages, record the baseline for AI-referred sessions before a big sales period so the effect can be judged afterwards.
What this means for Thai marketers
Adobe's numbers are from the US, but the direction is relevant to Thai retailers heading into 11.11 and 12.12. The practical lesson is about measurement and content, not about copying a US percentage. Thai online stores can segment AI-referred sessions in their own analytics to see whether the pattern holds for their customers, and can review whether product, FAQ and returns pages explain things clearly enough for an AI assistant to quote them. That work sits at the overlap of generative engine optimization and ecommerce marketing, and it depends on a clean GA4 setup that can actually tell AI referrals apart from other traffic. Relevant Audience covered Adobe's earlier holiday forecast at the start of October; this report adds September transaction data and the content-readability findings.
FAQ
How much more revenue do AI-referred visits generate?
Adobe reported that AI-referred visits to US retail sites produced 83% higher revenue per visit than non-AI traffic in September 2026. Adobe called it the largest gap it had observed this year and forecast that the advantage would double over the holiday season.
Does this data apply to Thailand?
No, the data covers US retail sites only. Adobe published no figures for Thailand or Southeast Asia in this report, so Thai retailers should measure AI-referred sessions in their own analytics before assuming a similar gap.
What share of retail traffic comes from AI?
Adobe did not say in this report. It published growth rates and comparisons between AI and non-AI visits, but not AI's share of total visits, so the revenue-per-visit premium cannot be turned into a total revenue figure from this source.
What is the Citation Readability Score?
It is an Adobe metric that measures how much of a web page's content AI can understand and cite. The average US retail homepage scored 63% in September 2026, Electronics led categories at 79%, and General Merchandise homepages scored 40%.
Which pages does AI read best, according to Adobe?
Adobe says AI reads explanatory content best, such as product details, buying guides, FAQs and return policies. In Electronics, customer service pages with setup guides, troubleshooting and detailed specifications scored 78%.
Adobe's report is a strong signal that AI assistants are sending shoppers who are close to buying, but it is a US signal with no share-of-traffic figure attached. The useful response is to measure AI referrals in your own data and make your most useful pages easy for AI to read. If you want help with either, talk to the Relevant Audience GEO team.
