Ranking well in Google's product carousel no longer means you will be shown the same way in AI Mode. New data from Productrise suggests Google can display different products, different sellers, and different prices for the same shopping query, even on the same day. For ecommerce teams, that changes the brief: AI shopping visibility is now its own battlefield.
The practical issue is simple. AI Mode shows a much shorter shortlist than the classic carousel, so the seller and price shown first can shape the buyer's decision before they compare alternatives. If your brand assumes strong free-listing visibility automatically carries into AI Mode, this dataset says otherwise.
The clearest finding is divergence. According to Productrise, which tracked more than 2 million product listings across more than 100,000 regular search results and AI Mode responses from August 9 to August 31, 2026, Google's two shopping surfaces often did not agree on what to show for the same search. The analysis compared product-focused queries in the US and UK and matched products using Google's product ID.
That methodology matters. On the regular SERP side, the comparison only looked at the Popular products carousel. On the AI Mode side, it looked at the product cards shown inside the answer. In other words, this was not a vague comparison between two different search moments. It was a same-query, same-day check of what buyers actually saw first.
| Metric | Popular products carousel | AI Mode | Why it matters |
|---|---|---|---|
| Average products shown when both views returned products | 27.8 | 3.9 | AI Mode creates a much tighter shortlist. |
| Daily product overlap | Baseline | 1.28% of carousel products also appeared | Carousel visibility rarely carried over into AI Mode. |
| Different lead seller on matched products | Not applicable | 49.6% | The merchant shown first often changed. |
| Different lead price on matched products | Not applicable | 38.1% | The visible offer often changed too. |
| When prices differed | Not applicable | AI Mode was higher 68.4% of the time | Buyers could be anchored to a more expensive offer. |
There is another number ecommerce teams should not ignore. Across matched products, including cases where the listed price stayed the same, AI Mode prices were on average 21.6% higher than carousel prices. Productrise also noted that when AI Mode showed the higher price, the median difference was 22.2%, while extreme outliers pushed the average gap higher. When AI Mode was cheaper, the median difference was 7.8%.
That does not prove Google is intentionally favoring expensive offers. It does prove something less dramatic and more important: the buying context a user sees in AI Mode is often not the same one they would see in the classic Shopping-style module.
The short answer is that a shared data source does not guarantee identical selection logic. Google said it has not verified the report, but it also said that "all shopping results on Google Search ... are powered by the same data source: our Shopping Graph." That means the raw product universe may be shared while the presentation layer and ranking choices still differ.
Google's own help documentation explains that the Shopping Graph is used across Search results, ads, YouTube, and generative AI features, drawing on product names, descriptions, prices, images, and reviews supplied by brands, retailers, and other content providers. The important nuance is that a common graph is not the same thing as a common interface. One surface can still pick a different offer, a different merchant, or a different subset of items from the same pool.
That is especially plausible when one experience shows nearly 28 products and the other shows fewer than 4. A large carousel can afford to expose breadth. An AI answer has to compress the choice set. Compression changes ranking pressure. It forces Google to decide not only which product is relevant, but which exact offer deserves the lead position inside a much smaller recommendation block.
A simple example makes this real. Imagine a shopper searches for a specific pair of running shoes. In the carousel, they may see multiple merchants and several price points at once. In AI Mode, they may instead see one lead card from a single seller. If that seller is different, or the card reflects a different condition, bundle, or freshness state, the buyer's first impression changes even if the underlying product identity is technically the same.
Productrise founder Hugo Huijer wrote that "the cheapest price is less of a factor in AI Mode." That is an interpretation, not a proven rule, but it fits the pattern the study found. It also fits what many AI interfaces do by design: simplify the comparison step for the user, even if that means showing fewer alternatives up front.
There is one more nuance that stops this story from becoming lazy clickbait. Clicking a product on either side can open a panel with other sellers and prices. So the first-listed offer is not the only offer available. Still, most users respond to the number they see first. If AI Mode changes that number or that seller, it changes the commercial reality of the SERP.
It changes the measurement model first. For ecommerce brands, this is now a Generative Engine Optimization (GEO) problem as much as a product-feed problem. Your AI search visibility can diverge sharply from your carousel visibility, even when Google is pulling from the same broader commerce infrastructure.
That is the key strategic takeaway. Classic free listings still matter. Merchant Center hygiene still matters. Feed completeness still matters. But the Productrise numbers suggest they are no longer enough as a proxy for what AI Mode recommends. If only 1.28% of carousel products also appeared in AI Mode on the same day for the same query, then a good carousel rank is a weak predictor of AI exposure.
This also exposes a reporting gap. Google's AI performance insights in Merchant Center help brands understand discovery across AI Mode and AI Overviews through metrics like share of voice, products showing, search intents, and missing product attributes. Useful, yes. Complete, no. As the documentation and the study together make clear, those reports do not tell you which seller or which exact price AI Mode put in front of the user.
That is why this topic matters beyond technical SEO. It reaches merchandising, pricing, feed ops, and brand control. If the visible merchant changes almost half the time for matched products, then AI Mode is not merely reflecting your catalog. It is shaping a recommendation moment.
We made a related point in our piece on Google turning Merchant Center into an AI shopping visibility dashboard. Native platform reporting is useful for directional insight, but it is still not the same thing as watching the actual answer a shopper receives. The interface matters because that is where buyer perception forms.
There is also a competitive angle. In AI shopping, recommendation and attribution can split apart. A brand may be relevant, yet the AI layer may foreground a different merchant, a different reseller, or a different evidence source. That is why ecommerce teams increasingly need to treat product visibility, merchant visibility, and commercial messaging as separate layers rather than one bundled SEO outcome.
The most important lesson here is not that AI Mode might show a higher price. It is that most ecommerce teams still lack clean observability into how AI shopping experiences frame their products. Merchant Center can tell you some things about AI discovery, and classic SEO tools can tell you some things about rankings, but neither one fully answers a basic commercial question: what did the buyer actually see first?
This is where BotRank's AI Visibility feature becomes genuinely useful. It lets teams run repeatable prompts across AI systems and track how brands appear over time, rather than guessing from occasional manual checks. For cases like this, pairing that with Source Analysis is even more valuable, because it helps you inspect which pages and domains AI systems appear to lean on when forming the answer. That will not magically fix seller or price mismatches. What it does give you is a way to separate three problems that too many teams still lump together: feed quality, AI answer visibility, and recommendation evidence. Teams weighing which platform to use for this can read our best GEO tool comparison and our top 5 GEO tools guide.
The answer is to operationalize comparison, not panic. This study does not mean AI Mode is broken. It means the old shortcut of using carousel performance as a stand-in for AI shopping performance is no longer safe.
Notice what is not on that list: random GEO hacks. This is not about gaming a prompt. It is about making the AI-facing version of your product information as consistent, trustworthy, and monitorable as the SEO-facing version.
The data is useful, but it is not a universal truth for every shopping query on Google. Productrise itself left important caveats in plain sight, and smart teams should keep them in view.
That nuance matters because the wrong reaction here is outrage theatre. The better reaction is operational skepticism. Treat the findings as a warning that AI Mode deserves direct monitoring, not as proof that every AI shopping result is worse for users or unfair to merchants. If you want that direct monitoring in place quickly, see BotRank pricing to pick the plan that fits your catalog.
No. In this dataset, only 1.28% of products in the carousel also appeared in AI Mode for the same query on the same day. That makes carousel ranking a weak indicator of AI Mode visibility.
The most likely explanation is that AI Mode can choose a different lead seller or a different offer context from the same product graph. That does not always mean bad data, but it does mean the buyer's first impression can change.
Not by itself. Merchant Center's AI reporting helps with visibility trends and share-of-voice style reporting, but it does not expose every seller or price choice shown in AI Mode. You still need direct observation of the answer layer.
Start with the basics that affect trust and comparability: price consistency, complete product attributes, accurate condition and availability data, and recurring checks on your top queries. Then build a repeatable monitoring workflow so AI Mode stops being a blind spot.
The takeaway is blunt: stop treating Google AI Mode as a mirror of the Shopping carousel. Audit your top product queries in both views, track where sellers and prices diverge, and build a process that turns those gaps into action. If your team wants a faster way to measure that new layer of ecommerce visibility, BotRank gives you the monitoring and diagnostic workflow to do it without flying blind.
After nearly 15 years in digital strategy on the client side (including 10 years at Olympique Lyonnais, where he was notably in charge of SEO).
Florian co-founded BotRank.ai in 2025, the GEO (Generative Engine Optimization) tool used by more than 2,500 companies to manage their visibility in AI-generated search results. He writes regularly about GEO and AI Search.
