Adobe's AI Catalogue Tool Reveals a Billion-Dollar Retail Blind Spot

Adobe launched Adobe Catalogue Agent for Adobe Commerce on 27 July 2026, a tool built to make product catalogues readable by AI shopping assistants. It arrives alongside Adobe data showing AI-sourced traffic to US retail sites grew 125% year-on-year between April and June 2026, while the average product page scores just 66% on machine readability. In short: retailers are getting far more AI traffic than a year ago, but most of their product pages remain half-invisible to the AI systems sending that traffic.

What changed?

Adobe has made product discovery on large language model (LLM) surfaces available inside Adobe Commerce, centred on the new Adobe Catalogue Agent, according to Stellagent. The tool takes structured product information already held in the Commerce catalogue and renders it as a machine-readable layer on product detail pages.

The launch is backed by fresh figures from Adobe Digital Insights. AI-sourced traffic to US retail sites grew 125% year over year from April through June 2026, yet individual product pages score just 66% on machine readability, the lowest of any major page type, per Adobe's own analysis. That means roughly a third of the content on the pages that carry actual purchase intent, price, stock, specs and reviews, is effectively invisible to the AI tools now driving a growing share of shopping traffic.

This follows a longer run of Adobe data tracking the same trend. Traffic from AI sources to US retail sites grew 393% year over year in the first three months of 2026, and AI-referred traffic grew 138% year over year in May 2026, up 1,324% (more than 14x) since October 2024, when Adobe first began tracking these referrals, according to Digital Commerce 360. During the most recent holiday season, the figure was up 693%, and the series suggests growth is decelerating in percentage terms while absolute volume keeps compounding, per VMblog's coverage.

Crucially, this isn't traffic for traffic's sake. AI traffic converted 42% better than non-AI traffic in March 2026, a new record high, according to Marketing Week. Once a shopper lands on a US retail site from an AI source, engagement is 12% higher, with visits lasting 48% longer and covering 13% more pages.

Why does this matter for your business?

If you sell anything online, this is a live money problem, not a future one. More AI-driven demand is arriving at your door, but a large share of your pages can't be read properly by the systems sending it. That means you could be losing sales to competitors with cleaner product data, even if your product is genuinely better.

The gap isn't spread evenly. Cosmetics and electronics led on overall AI readability in May, with cosmetics sites at 63% and electronics at 56%, per Adobe's data. Product pages fare worse than other page types across the board, and that matters because product pages are precisely where AI shopping assistants need to pull specifications, pricing and availability to make a recommendation.

This isn't a US-only issue either. New Adobe data shows major portions of UK retail websites are not entirely readable by machines, limiting their visibility across AI search results, according to VMblog. Adobe's AI Content Visibility Checker puts the average UK product page at the same 66 out of 100 score seen in the US market, suggesting this is a structural problem across markets rather than a quirk of one region's web practices.

The stakes are rising because consumer habit is catching up with the technology. In Adobe's survey, 39% of consumers say they have used AI for online shopping before, and 85% of them say it improved their experience, per Marketing Week. That's a strong signal that AI-assisted shopping is becoming a habit rather than a novelty, and habits compound. Retailers who fix machine readability now are building an early lead that gets harder to close later; those who don't risk becoming structurally invisible in a channel that's growing every quarter.

What should you do now?

Start by finding out where you actually stand, rather than assuming your site is fine because it looks fine to a human visitor. Adobe built its own diagnostic tool for exactly this reason: the AI Content Visibility Checker shows how much of a webpage is actually visible to AI search tools, comparing an AI agent's view of the page against a user's view and highlighting content that stays hidden to agents.

Beyond running a check, the practical priorities for any retailer or brand are straightforward:

• Structure your product data properly. Attributes, use cases and compatibility details need to sit in a format AI systems can parse, not just in images, embedded widgets or JavaScript-rendered elements a crawler can't read.
• Prioritise product and category pages first. These are consistently the weakest page types on machine readability, and they're also the pages that directly influence purchase decisions.
• Treat this as ongoing maintenance, not a one-off fix. AI traffic volumes are still climbing quarter on quarter, so a readability gap left unaddressed now compounds as more of your prospective customers arrive through AI channels.
• Check your visibility across AI platforms broadly, not just retail-specific tools. AI shopping assistants pull from general-purpose models as well as dedicated commerce tools, so the same structural fixes that help Google's AI Overviews or ChatGPT cite you correctly also help retail-specific AI discovery.

This is also a good moment for any brand, retail or otherwise, to check where it currently stands with AI systems more broadly: Sited's free audit at https://sited.online gives a quick read on whether AI tools are actually citing and recommending your brand today, which is a useful baseline before you start structural fixes.

Frequently asked questions

What is Adobe's AI Content Visibility Checker?

It's a free browser tool from Adobe that analyses a webpage and reports what percentage of its content large language models can actually read. It's a standalone diagnostic tool that works directly in your browser, with no setup or Adobe licence required.

Why do product pages score worse than other page types?

Product pages typically rely on dynamic elements, embedded specification tables, image-based details and JavaScript rendering that can be harder for AI crawlers to parse consistently, compared with simpler content pages like blog posts or FAQs.

Is this only a problem for Adobe Commerce customers?

No. The underlying readability issue applies to any retail website, regardless of platform. How thoroughly attributes, use cases and compatibility are described is a shared competitive condition, whether or not you use Adobe Commerce.

How much has AI traffic to retail sites actually grown?

The exact figure depends on the period measured, but the trend is consistently steep. Traffic from AI sources to US retail sites grew 393% year over year in the first three months of 2026, while more recent data puts April-to-June growth at 125% year-on-year, showing continued strong growth even as the rate moderates.

Does fixing machine readability actually improve sales?

Adobe's data suggests it should, given that AI-referred traffic already converts significantly better than other channels. AI traffic converted 42% better than non-AI traffic in March 2026, a record high, so improving how AI systems read your pages is likely to expand your share of an already higher-converting audience.