AI shoppers convert better than other visitors on US retail sites. Can AI read your homepage?

What Adobe reported on 19 August

Adobe's July 2026 data says AI referrals to US retail sites are still growing and still convert better than other visits. AI-referral traffic rose 62% year on year in July, based on more than 1 trillion visits to US retail sites tracked in Adobe Analytics.¹ Adobe published the figures on its blog on 19 August. Against October 2024, when Adobe started tracking the channel, AI referrals were up 1,219%.¹

The quality figures matter more than the growth. In July 2026, AI-referred visits converted 60% better than non-AI traffic.¹ Adobe said July was the eleventh straight month in which AI traffic outperformed other traffic on conversion.¹

A disclosure the reader needs: Adobe sells analytics and AI visibility tools, so it has an interest in this channel looking large. Its sample is still the broadest public one for US retail.

How the July numbers compare with spring

Growth is slowing while the conversion premium widens. Adobe's first-quarter report put AI-sourced traffic to US retail sites up 393% year on year in Q1 2026, with AI visits converting 42% better than other traffic in March 2026.² By May 2026, Adobe's figure for US retail was 54% better conversion, with AI visits worth 53% more.³ July's 60% is the highest of the three readings.¹

Falling growth rates are what happens when a base gets bigger. A channel that was tiny in mid-2025 could triple easily; doubling a larger base is harder. Read the 62% as a channel settling in, not stalling.¹

Behaviour on site points the same way. In July, visitors from AI platforms engaged 14% more, spent 59% more time on site, bounced 33% less and added items to their cart 28% more often than other visitors.¹ Those are the habits of someone who has already compared options in a chat and arrives to check one.

Shopify told a similar story from the merchant side. On its earnings call in August 2026, it said AI-driven traffic and orders to its stores had tripled year on year in Q2 2026, while search still made up roughly a third of storefront sessions.⁴ We covered that call in our note on Shopify's AI orders.

Why does AI traffic convert better?

Mostly because of who arrives. A shopper who clicks a link inside an AI answer has usually asked a specific question, read a summary of options and picked one. The research happened before the visit. That is selection, and Adobe's figures do not separate it from any effect of the AI answer itself.

Our view: the premium is real for planning purposes, but it is a correlation. It tells you AI-referred visitors are worth having. It does not tell you that sending more of them will lift your average conversion rate by the same margin, because the next tranche of AI visitors may be earlier in their journey.

It also hides variation. Adobe's figures are averages across very different retailers, and categories behave differently.¹ Measure your own AI referral conversion in analytics before you use any of these figures in a business case.

The part most retailers can fix: 39% of homepage content

The same release carried a finding retailers can act on this month. Adobe's AI Content Visibility Checker scores how much of a page a large language model can read. Across a broader cohort of US retail sites in July 2026, homepages scored 61% on average, so about 39% of homepage content was not easily machine-readable.¹

Adobe widened the cohort between April and July without saying by how much, so the July average is not comparable with its earlier readings. Adobe also scored categories, but on overall readability across homepages, product pages and corporate blogs, not homepages alone.¹ The split is still useful.

Adobe linked the leaders' scores to consistent, structured product content. Our view: the gap is usually mundane. Promotions sit in image banners, category names live in menus built by script, and shipping or returns terms hide behind tabs that only render after a click.

A homepage check you can run in an hour

Look at your homepage the way a crawler does, then fix what is missing. These five checks cover most of what we find.

1. Load it without JavaScript. Disable scripts in your browser or fetch the raw HTML. Every category, promise and policy a buyer needs should still be there as text.
2. Read the banners. Any offer, delivery threshold or guarantee that exists only inside an image is invisible to most AI systems. Put the words in HTML next to it.
3. State the facts answers repeat. What you sell, who ships where, how returns work and how to contact you, in plain sentences near the top of the page.
4. Link to the pages that hold detail. Category pages, the returns policy and sizing guides should be one crawlable link away, not buried in a script-driven menu.
5. Keep prices and stock honest. If a product is shown with a price, the server-rendered page and your feed should carry the same one.

For the product-level version of this work, our guide to why ChatGPT does not recommend your products covers feeds, retailer listings and the sources the shopping answers draw on. Our page for e-commerce and D2C brands sets out the prompts we measure for stores.

What to take from the July release

AI shoppers are a small, valuable and growing group on US retail sites, and many homepages still hide basic facts from the systems that send them. The first part of that sentence is Adobe's to measure. The second part is yours to fix.

Our view, and the order we follow at Sigzen AI: do the readability work before you spend on content or placements meant to win AI mentions. If the engine that recommends you lands its user on a homepage it could not read, you are paying for a visit it was not able to explain.