Search for a product on Google today, and you’ll be directed to two completely different results pages, depending on which button you click. The standard search box returns the usual results page, with a carousel of products at the top. Click the AI Mode button, located next to the search box on Chrome’s address bar or on a tab next to your results, and the same query returns a written response, with products “featured within” the response.
According to new research from product discovery platform Productrise, which tracks products’ visibilty on Google Shopping, the choice between the two options may not be as straightforward as it seems. An analysis of over 24 thousand products found that, on average, the products featured in AI Mode are priced 21.6% higher than the same products in the traditional search results. The findings have since been highlighted by Futurism, Search Engine Journal, PCMag, Digital Trends, and PPC Land, reopening a question that’s been building for a while: who is AI Mode actually optimizing for?
What Is Google AI Mode, And Why Does It Matter For Shopping
For those who have not fallen under its spell, AI Mode is Google’s new conversational search interface, which exists not in place of the traditional results page, but alongside it for now. This mode offers something a little different than the usual ten blue links and a smattering of ads and shoppable carousels. When activated, queries are met with a passage of text, often including aggregated information from several sources. Queries with a clear commercial intent get cards as answers, similar to the carousel of ads and shoppable products found elsewhere on the results page, but with a different production pipeline.
Artificially intelligent search has had a slow and steady rollout over the course of the last year. Google has seen fit to both install it for all Search Labs users who opted in and has begun to bake the feature more directly into the core search experience. The company has also made changes to the layout of the main page and the browser UI to make accessing AI mode both easier and, perhaps, more intrusive, replacing some older elements that once prompted or assisted in discovering the feature. Last month, Google changed the homepage to offer not just a “Search” button, but a series of AI-first features and modes. Several outlets covering the story for the recent article about Google’s pricing for AI Mode noted this as part of a larger trend in which Google has been pushing users toward its AI features, even if they have not opted in or otherwise expressed interest in the technology.
In the context of shopping, this matters far more than it would for other verticals. While the presence of a different set of sources behind an AI-generated answer might be seen as a quality control issue for informational queries, for commercial ones, it represents a fundamental change to the purchasing power of the searcher. For shopping searches, the difference matters more because the results can directly affect what someone pays. If AI Mode chooses different products, sellers or prices, the result can have a real impact on the shopper’s final purchase.
What Productrise Actually Tracked
Productrise conducted its analysis over the course of 23 days, from August 9 to August 31, 2026. Over this period, and across more than 100,000 searches, the company queried more than 2 million product listings using identical product-focused shopping searches across both Google’s standard search results and its AI Mode on the same day. The analysis was performed across the United States and United Kingdom, and prices were reported in dollars (US) and pounds (UK), rather than converted to a single currency.
For the purposes of its analysis on traditional searches, Productrise restricted its consideration to the “Popular products” carousel only, the strip of shoppable listings that appears near the top of a search results page, for comparison with the results returned via AI Mode, where it looked at every product card included in the AI response. For the purposes of direct comparison between the two, Productrise matched products between the two surfaces based on Google’s stable product identifier (where available), the same internal reference Google uses to identify a particular item across its Shopping Graph and for the purposes of tracking a given item’s availability and performance across different retailers for the same query on the same day.
That distinction matters. By using Google’s own product identifier to match listings on both sides, Productrise could compare the same product rather than relying on similar titles or images. That makes the price comparison much more meaningful.
In calculating its headline pricing comparisons, the company took the lead offer on either side of the equation (in other words, whichever seller’s listing appeared in the top position), and calculated the percentage difference between the AI Mode price and the traditional search price for that particular product. A separate calculation was then made, looking at every priced listing on either side, rather than just the matched pairs, in order to compare the overall pricing mix between the two surfaces for the purpose of calculating what each was offering shoppers in terms of price ranges.
Interestingly, this is Productrise’s second study on AI Mode. The company’s previous release in July 2026, found that AI Mode tends to return roughly 95% fewer product listings than standard search on the same queries, and this new study builds on that, asking a follow-up question of sorts: when AI Mode does show a product, what does it cost?
Why The Product Identifier Matters
A lot of the credibility of a study like this rests on how carefully it defines what counts as “the same product.” Two different product listings can be presented with the same image and title, roughly comparable prices, but different shipping options, and therefore different overall prices, from different merchants. Sloppy definitions of “the same product” risk producing interesting-sounding results from loose comparisons between similar goods, while tighter definitions yield a more useful, relevant study.
Productrise was able to avoid these issues by defining the same product as entries that used the same Google stable product ID – a unique, internal identifier in Google’s Shopping Graph, which is their internal database for all products indexed across Google Shopping and AI Mode. By virtue of both the traditional carousel results and the AI Mode results coming from Google Shopping – and therefore using the same product database – a stable ID is one of the few ways to be reasonably confident that the same product is being compared between two different product listings. Productrise has documentation on their methodology for their daily tracking product, which confirms that they use this stable ID to determine if the same product has shown up on both the traditional Google Shopping carousel and on AI Mode. In short, this is the most reasonable methodology to define products shown on both interfaces as being “the same.”
That makes the 1.28% overlap particularly striking. In most searches, the products shown in standard search and AI Mode were completely different, even when the query was identical.
The Headline Number: 21.6% More Expensive
The key takeaway is simple. For products appearing in both traditional search and AI Mode, for the same query, on the same day, the price AI Mode showed was on average 21.6% higher than the price shown in the standard search carousel.
The difference is large enough to matter. On a $100 product, for example, a 21.6% gap would mean paying about $21 more for the same item.
Independent SEO Consultant Brodie Clark (who reviewed the findings for Productrise) noted that it’s not an entirely clean apples-to-apples comparison, since AI Mode can contain a higher-priced item within the grid result itself (complicating a top-level comparison), but also made a point that hits right at the heart of why this study is worthwhile. In traditional product listings, it’s usually the cheapest item that takes top placement and best presentation, but that pattern appeared to break down in AI Mode, where price did not appear to be a deciding factor in what got recommended.
Senior SEO Strategist at Break The Web Katelyn Geary put a blunt spin on the implication for shoppers. AI search was supposed to help by taking the hassle out of having to compare prices across multiple browser tabs, she said, but if it’s secretly pushing pricier inventory by default, it hasn’t actually saved anyone any work. The convenience is real, but shoppers may be giving up some price visibility in exchange for it. Someone who accepts the first recommendation may never realize that the same product was available for less elsewhere.
Finding Two: The Broader Price Gap Is Even Bigger
The matched-product comparison is the cleanest test that Productrise ran, but it only tells part of the story, as it only covers products that appear on both surfaces. To get a fuller understanding, Productrise compared every priced listing on either side, including products that appeared on only one of the two surfaces
On that broader measure, the median price in AI Mode was $149, compared to a median of $100 in traditional search. That’s AI Mode listings run roughly 49% higher than what standard search typically shows for the same category of query.
Part of the explanation is structural. AI Mode simply displays far fewer products than standard search to begin with. Across the study, AI Mode made up just 12.3% of all the listings Productrise tracked, showing an average of 3.9 products per search compared to 27.8 in the traditional carousel. AI Mode is working with a much smaller pool of products. That means its selection criteria can have a much bigger effect on the prices shoppers see.
Finding Three: The Two Surfaces Barely Show The Same Products
Even when a shopper runs the identical search on both traditional search and AI Mode, on the same day, the two surfaces rarely agree on which products to show. Only 1.28% of products which appeared in position on the traditional search carousel also appeared in the AI Mode for the same search on the same day, according to the study by Productrise. This observation aligns with the data from the company’s July analysis, which revealed that AI Mode yields roughly 95% fewer listings than traditional searches. The latest study serves to corroborate this trend with an additional finding – in cases when there is overlap, the shared products are weak indicators of what AI Mode is likely to recommend.
Jamie D’Alessandro, SEO Manager at JAKALA, noted one particularly intriguing observation from the analysis. The most common overlap occurred when neither surface was able to detect a relevant listing. However, in cases when both surfaces did respond to the same query, the overlap was uncommon, and the products which appeared on both tended to be contrary to Google’s established ranking principles, which prioritize cheap listings most prominently, followed by enhanced visual presentation. The same principles failed to apply to AI Mode.
Another notable observation, which D’Alessandro added was not captured by the study, but was apparent from the screenshots – AI Mode appears far more likely to promote a brand’s website over a marketplace listing. If this observation holds true, it suggests that AI Mode’s recommendation is far more susceptible to elements which are directly controlled by the brand, such as the presence of Product Highlights or the depth of product data provided via the product feed, rather than the inherent characteristics of marketplace listings, such as being the cheapest option available.
What The Price Gaps Actually Look Like
After numbers and percentages, the most crucial information is a contextual description of what the discrepancies found by the researchers actually meant for the shoppers. Fortunately, Productrise has provided a few vital examples that illustrate its findings concerning the differences between search results on the standard Google and the AI Mode.
First of all, there was a case when the same product was shown in the same position on both surfaces; however, the one in AI Mode was stated to be 26 percent more expensive. There were no changes in the search terms or the product; the only difference was the mode of search, which resulted in a significant difference in prices. Another example illustrated that changes in sellers could also be considered a discrepancy. While in the standard search engine, the product was promoted by a particular seller at a lower price, Google AI search suggested a different seller with a higher price (approximately 19 percent). This is the pattern Callum Lockwood flagged as more concerning than the price gap itself: AI Mode isn’t just re-ranking the same offer, it’s choosing a different one entirely based on its own criteria.
The third example described by the author concerns a situation when a used product was promoted in standard search while a new one of the same product was promoted in AI Mode. Productrise representatives indicated that such cases impacted their average price-gap statistics in favor of higher discrepancies as, in theory, a used product could be compared to a new one. The difference between them is significantly greater than the average price-gap statistics reported by the company.
However, not all of the cases illustrated the AI search’s deliberate promotion of different higher-priced offers. According to the report, 68 percent of all matches illustrated no discrepancies. In some cases, not only the offers but sellers remained the same; however, the description of one product was slightly modified to fit the requirements of Google AI search. Thus, the same offer appeared in two different positions with slightly different titles. Furthermore, as stated by the report, in one case, one particular offer was promoted in the eighth position in standard search while in the first position in AI Mode search.
Taken together, the examples show that the price differences are not limited to one unusual case. Sometimes AI Mode showed a different seller, sometimes the same product at a different price, and sometimes there was little or no difference at all. The important point is that shoppers cannot assume the first AI Mode result is the cheapest available option.
Finding Four: When Prices Disagree, AI Mode Is Usually The Pricier One
Not every matched product has a price discrepancy. Productrise discovered that prices disagreed on 38.1 percent of products that were matched. But on the occasions when the AI Mode had the pricier surface, that was the case for 68.4 percent of disagreements.
When AI Mode appeared as the pricier surface, the median markup was 22.2 percent. The average gap was far larger, at 88.5 percent, but Productrise warned that this figure is skewed upward by outliers, such as a used or refurbished product that appeared in traditional search being matched against a brand-new version of the same product priced higher in AI Mode.
Tristan James, an SEO at EE, said the finding is particularly relevant while shoppers are dealing with continued cost-of-living pressures. If AI Mode regularly puts more expensive products at the top, he argued, the convenience of a simpler shopping experience may not be worth much to consumers.
He also suggested that brands should regularly check what AI Mode is saying about their products. That includes comparing the pricing information shown by AI Mode with the pricing data available elsewhere, rather than assuming the two will always match.
For James, the pragmatic takeaway was that brands now needed to treat “fact-checking” what AI Mode said about their products as an ongoing task, not a one-time exercise. He said he intended to pay closer attention to the intersection of SEO and what is sometimes known as GEO, or generative engine optimization, to ensure alignment of every pricing data source a brand had behind a single consistent number.
Finding Five: When AI Mode Is Cheaper, The Gap Is Smaller
The study also examined the opposite case. In the instances where AI Mode showed the lower price, the median difference was 7.8% cheaper, and the average was 11.9% cheaper. Both lower than the markups observed in the cases where AI Mode was the more expensive surface, supporting the hypothesis that divergences in pricing in favor of AI Mode tend to be more often and significantly positive than negative.
Finding Six: A Different Seller Nearly Half The Time
Price was not the only parameter changing between surfaces. On instances where the same product appeared on both traditional search and AI Mode, the top seller was different 49.6% of the time, almost a flip of the coin. Productrise pointed out that in such cases, changes in price and seller tend to be related, meaning that AI Mode is rarely simply re-ranking offers from the same retailer at a different price point, but rather is frequently showcasing an entirely different seller for that particular item.
Callum Lockwood, Director of Organic Search at Re, said the seller changes may be more important for retailers than the price differences themselves. Nearly half of the matched products were shown with a different seller in AI Mode, suggesting Google may be selecting different offers rather than simply displaying the same products differently.
“Feed data quality and stock availability are obvious drivers,” Lockwood said. “But AI-powered shopping surfaces also look beyond a merchant’s own product feed. Customer reviews, third-party citations, press coverage, and other signals from across the web can all play a role in how merchants are selected.”
Google’s Response
Google was given an opportunity to respond to the study’s findings before Futurism published its report on September 2. Their statement, as reported by multiple outlets covering the study, did not explicitly reject the validity of the methods used but did not directly confirm the figures, either. Google stated that it had no independent confirmation of accuracy of the findings in the report, and stressed that every shopping results in Google Search, whether accessed via AI Mode or the standard results page, were pulled from the same repository of information, which the company calls its Shopping Graph. Google added that shoppers could click into individual product listings and see comparisons between different sellers and choose the option most beneficial to them.
The last point is a technically accurate description of how both AI Mode and shopping searches without it work in tandem with the main search function: clicking into a listing would open a secondary panel with other available options for the same product, priced differently by other retailers. However, several of the SEO commentators covering the study pushed back on how effective such a safeguard mechanism was, arguing that most shoppers would not click through to another page. When presented with a recommendation, on either interface, that presents itself as the most beneficial option, the expectation of getting the best price is justified, without needing to navigate away from the listing to other retailers, since the first option presented has already been deemed the most beneficial
Jeff Collins, an independent SEO consultant commenting on the findings of the study, remarked on this as anchoring, where the tendency for a price to be seen as the best is informed by the first number a shopper sees. In the context of AI Mode and Productrise’s data, this means that Google has an incentive to push for conversions, moving the shopper from asking a question to actually purchasing a product, and a shorter buying cycle aids that greatly, remarked Collins. A listing with a few products, suggested by the AI, is infinitely more effective at moving the buyer than navigating a longer list of 20, he remarked, noting that the format itself encourages anchoring by virtue of the AI results being wrapped in more extensive, explanatory text, compressing the information garnered from 5 separate open browser tabs into a single conversation. If this trend was a permanent fixture of Google Shopping, he expected buying cycles to be shortened overall and brands to invest more heavily in ensuring that their product data and content was optimized to appear in the Google Shopping results at all stages of the buying cycle, not just the final purchase.
The Surveillance Pricing Question
In the days following the release of the study in question, some commentators took the findings as indicative of a separate but related phenomenon: surveillance pricing, the practice of using a shopper’s personal data to adjust prices presented to them in real time, based on such data as browsing history, location, or purchasing patterns. US lawmakers have already begun scrutinizing the practice more generally, not limited to Google or AI search specifically.
It is worth being precise about what the Productrise study claims or does not claim in this regard. The study looks at what different Google surfaces present for the same query run at the same time, not whether Google is personalizing the prices it shows to individual shoppers based on some of their data. Productrise has not claimed that, and the study does not attempt to prove that. The study only provides a concrete number that reflects a pattern that many shoppers may have already suspected on an anecdotal level: that the same product can have two different price tags depending on which version of Google search one runs into. Whether or not this pattern has anything to do with surveillance pricing is a related but arguably simpler question that does not require any personal data to be involved at all: why is the same product consistently more expensive on one Google surface than another, and who benefits from sending shoppers to the more expensive one by default?
How This Fits Into Google’s Longer History With Shopping Complaints
This is not the first time Google’s shopping results have been accused of containing unfairly biased information toward certain sellers and, through them, shoppers. Google’s various shopping products have been the subject of regulatory and antitrust scrutiny around the world over the last decade. The previously mentioned issues concerned Google’s dominance in shopping search results; the complaints included that Google’s search results unfairly prioritized Google Shopping results and advertisements over alternative price comparison shopping sites and independent retailers. In essence, the criticized practice involved Google’s alleged ability to dictate who got seen and who did not in the search results snippet.
It should be noted that the Productrise study is not a legal complaint, nor does it allege any anti-competitive behavior on Google’s part. Instead, the recently published study compares the previously mentioned accusations to the new circumstances created by the LLM technology layering on top of Google Search. Effectively, the Productrise researchers asked the same questions that have been asked by the regulators and the public regarding Google Shopping’s propensity to return higher positions to Google’s own listings while replacing the concept of verticals in search with products and, in turn, prices.
The Productrise study does not provide any direct answers to the questions surrounding Google Shopping. However, as demonstrated by the product’s launch, there is an evident interest among the public and the regulators to scrutinize Google’s practices in the space. As such, the report resonated far beyond the niche audience of SEO professionals as it struck at the perceived core of Google Shopping’s dominance in the market. In the context of shopping results, the unfairness to shoppers is a known issue that has already drawn the attention of several regulatory authorities. In this regard, mentioning a specific value that has the potential to demonstrate Google’s alleged bias most effectively is not surprising.
How Shoppers End Up In AI Mode Without Realizing It
Part of the explanation for the large discrepancy in price lies in how easily, and frequently invisibly, a shopper can be nudged into the AI Mode. Google has been promoting the AI Mode as a major feature of its search engine throughout the past year, and there are several places within the search engine ecosystem where a shopper can access it without even having to type a query.
Google has placed a large AI Mode discovery CTA right on the standard search homepage, alongside the standard query bar. This allows the user to open AI Mode while preserving their original query. Chrome also exposes an AI Mode button on the address bar directly, allowing a search to be rerouted with minimal effort. Finally, there is an AI Mode tab on a standard results page, above the results themselves, asking the user to reroute their query through the AI interface.
As reported by PPC Land, who have been tracking the evolution of Google’s search product, the company began rolling out experimental AI Mode buttons within the home page search bar for Search Labs users in June 2025. The address bar button became available for Chrome users about three months later, in September 2025. Researchers outside of Google were able to confirm in July 2026 that the address bar button was routing queries to Google’s hosted AI models in the cloud, not the smaller Gemini Nano running on the user’s local machine.
The implication of all of this is that a shopper doesn’t have to seek out the AI Mode to begin with, only to realize they’ve been tricked into using it. With minimal extra effort beyond a standard search, the user can reroute their query to a different set of results, potentially with a different ranking algorithm and a different price. This is why most price discrepancies occur — because a standard search can be rerouted into the AI Mode with an extra click, but the query itself rarely changes, not giving the user any warning that they’ve switched to a fundamentally different set of results.
Hugo Huijer, founder of Productrise, believes that this places far too much power in the hands of Google. He argued that as Google continues pulling AI Mode closer to the core search experience, the first listing will act as an anchor, since most shoppers won’t click past it.
Huijer also believes that this evolution in Google’s search algorithm can serve as a wake-up call to brands that have little pricing flexibility. If the new AI-driven Google Shopping results do not prioritize lowest-price wins, he reasoned, then their only hope in competing with bigger, more aggressive rivals lies in maximizing the value of every element on their product listing beyond price.
Why This Matters For Everyday Shoppers
Take a step away from the SEO commentary and the main issue with this report is actually pretty simple. Shoppers trust that the best recommendation comes from the top of a search engine’s results, whether it be in terms of relevance, quality, or, very often, price. That trust is a largely reasonable assumption for traditional search results, where the cheapest listing for a given product tends to get the best placement and the clearest presentation.
According to Productrise’s data, that assumption may not translate to AI Mode, where the best recommendation appears to be significantly more expensive on average than the same query in standard search results. If most shoppers will never go beyond the first recommendation anyway, this amounts to a tax on convenience. You get a faster and more conversational shopping experience, but you’ll also be paying more for the privilege, without any indication in the UI that a trade-off is taking place
This does not mean that an AI Mode result will always be more expensive, and it is also worth noting that the study found AI Mode to be cheaper in a meaningful minority of matched cases. However, the trend of more expensive results far outweighs the minority of cases where shoppers would be charged less, as well as the margin of the price difference in either case.
These trends are large enough that it is no longer a safe assumption that an AI Mode listing will see the same value from the shopper’s perspective as a standard search would.
Understanding The Shopping Graph, And Why Google’s Defense Leans On It
Google’s rejoinder to the study is built on a single pillar – their assertion that AI Mode, standard search, and the regular results page all rely on the same pool of data, which is known as the Shopping Graph. There is reason to believe that this pillar, while close to being fully truthful, does not provide nearly enough information to support the intended conclusion.
The Shopping Graph is essentially Google’s internal database that aggregates product-related information from millions of stores, marketplaces, and retailers across the globe. It holds a vast amount of different kinds of data, ranging from basic product identifiers and pricing to seller information, reviews, and other structured data points. The information in the Shopping Graph is constantly updated as sellers update their listings in the system. Thus, when Google claims that all of their shopping tools use the same data, they are most likely telling the truth, as all of these services indeed rely on the same database.
However, having the same data set does not necessarily mean that the outcomes are the same, as a simple analogy with restaurants can demonstrate. A single restaurant with a single pantry can still prepare dozens of different meals based on the ingredients that it possesses, as different recipes call for different combinations of available products. By the same token, different ranking algorithms can extract different value propositions from the same data set, depending on their specific implementation. Standard search has had years of refinement, during which it has developed a certain set of priorities, such as price and value, that are preferred by Google and its customers. Thus, the cheapest product tends to rank highest within the standard search results, as it offers the broadest appeal. AI Mode, on the other hand, is presumably a newer tool, and its algorithm most likely relies on different principles, which means that it can prioritize different aspects, such as the degree of match between a product’s description and the query, the quality of its structured data and Product Highlights, or the strength of its reviews and seller information.
Prioritizing factors is not a sign of wrongdoing on Google’s part, as different buyers tend to have different priorities when shopping. Some will always go for the cheapest available option, while others might be willing to pay extra to secure a hassle-free refund or purchase a higher-quality product. The issue with Productrise’s study lies in the fact that it demonstrates that Google’s shoppers might not be aware of the differences in their search policies and that the company did not explicitly communicate these differences to its customers. While Google’s response to the study is technically factual, it essentially evades the subject, thus confirming the suspicions of many consumers.
What GEO Means For Brands Chasing AI Mode Visibility
Several of the SEO professionals who reviewed the paper mentioned a relatively new discipline that is developing around this shift, generative engine optimization, sometimes shortened to GEO, which is related but different from traditional SEO. SEO has traditionally focused on ranking factors for organic search results, but GEO looks at the different, more opaque set of signals that dictate whether an AI will chose to reference, recommend or cite a particular product, brand, or content within a generated result.
The Productrise study provides an early (if incomplete) data point as to what GEO might reward in a shopping context, specifically. If price was the dominant factor in what the AI Mode chose to recommend, the study would presumably have found AI Mode prices tracking close to or below standard search prices, and replicating the value driven pattern of the traditional carousel. Instead, things point the other way, indicating that whatever the underlying model is weighing in AI Mode, it is not simply defaulting to the cheapest available option as older ranking systems have done.
This creates both an opportunity and a risk for brands looking to decide where to invest: the opportunity, as Hugo Huijer and others identified, is that brands which could not compete solely on price may find AI Mode to be a more hospitable environment than the traditional price-driven carousel if they are able to invest in the richer product and reputation signals which seem to drive AI Mode selection. The risk is that this is an early and still-developing system; Google has not published detailed documentation on exactly what the AI Mode weighs in choosing what product and seller to recommend, meaning that brands investing heavily in GEO strategies today are to some degree working with inferred data and observed patterns like those in this study rather than a published rulebook which search marketers came to in traditional SEO for many of the traditional ranking factors.
For now, the most concrete and actionable step available to brands is the one several commentators pointed back to repeatedly: check for yourself. Run the searches a real customer would run, and examine what AI Mode is surfacing for those queries, comparing them to the traditional carousel and what your competitors are surfacing for similar queries. This will not reveal Google’s algorithm, but it will reveal discrepancies, such as a competitor’s listing outranking yours in AI Mode even as your product ranks higher in standard search, which are worth addressing regardless of cause.
What This Means For Brands And Retailers
For retailers and brands, the implications are more nuanced. The study’s finding that the primary listed merchant is different on nearly half of matched products means that a brand’s standing in the standard carousel has limited value in understanding its opportunities and risks in AI Mode. A brand may well be dominating the first page of normal Google results for a particular product, but if another seller (or even a competitor brand) is being selected more frequently by Google’s AI as the lead option in AI Mode, it has less ability to benefit from the traffic generated by the query.
This is a crucial point because Google’s Merchant Center provides an AI performance insights report that brands can use to understand how they are faring in both AI Mode and AI Overviews as a percentage of impressions. However, as the study underlines, that information only provides context about the overall presence. It does not show which merchant is being selected more often for a particular product or provide any information about the prices being shown. Thus, even if a brand is using the AI Performance Insights report to track its presence within Google’s AI results, it would have no insight from that report alone as to whether another seller was capturing more traffic in these AI-driven formats for similar products and at what price points.
The SEO experts who contributed their analysis for the Productrise study agree that brands should start thinking about manually checking what appears in AI Mode for their important products. In particular, they suggest that brands should perform these checks alongside their standard-search queries to understand if and how what appears in AI Mode differs from the traditional carousel. In doing so, they should not only be looking at whether the brand is appearing but also be evaluating the prices and merchants shown.
Notably, if Google’s algorithm for determining which merchant to highlight in AI Mode is relying on factors beyond price, it could have intriguing implications for search marketing. If elements such as the depth of a brand’s product feeds, the inclusion of structured data, customer reviews, and mentions elsewhere on the web, all play a role in determining prominence, brands that are heavy investors in these areas may well find that they are disproportionately rewarded within AI Mode compared to traditional searches. For some brands that are less able to compete on price with larger retailers and marketplaces, that could be an appealing prospect. At the same time, however, brands that have invested less in these areas (perhaps assuming that their superior pricing would be sufficient) may find that their presence in AI Mode is lower than their standard-search presence.
The Media Response
The study’s controversial claim that Google’s AI Mode delivers the same products at significantly higher prices than a standard search quickly gained traction after its publication, with multiple sources highlighting the discovery in an article titled “Google’s AI Mode Is Ripping Shoppers Off With Vastly Higher Prices, Study Claims.” Futurism’s coverage centered around the implications of the findings on Google’s push toward AI-powered search and away from its conventional results page design practices. Search Engine Journal, on the other hand, focused more on the mechanics behind the reported difference and the low level of product overlap between the two surfaces along with the frequent seller mismatches, considering the discrepancy only one of the supporting arguments. PCMag and Digital Trends also reported on the issue, with the latter outlet posing the question of how much Google’s defenders could twist the findings to spin a better light on the launch, as simply telling users to compare prices would be largely unhelpful, given the whole point of the AI Mode feature was to save them the trouble.
Both trade publications covering primarily SEO and SEM news and more mainstream technology sites geared toward non-specialists addressing a general audience weighing in on the report and its findings suggest that it has the potential to resonate with a considerably larger audience than just the SEO industry insiders, who are often the targets of such research. This is because the findings touch on a subject that is far more approachable than many technical nuances detailed in Google’s algorithm updates. A claim that a shopping assistant feature may, in fact, provide worse value than a standard Google search is easily comprehensible to any consumer-facing website user and, therefore, has a much broader appeal.
What Shoppers Can Do In The Meantime
None of this suggests AI mode should be avoided altogether, but the study does provide a simple, practical lesson for anyone searching through it: given the data, and the fact that the same knowledge panel showing alternate sellers appears on both surfaces, the safest behavior is to consider what the first AI product is recommending a price point, not a price floor. Clicking through to see the full list of sellers, or indeed just opening the same search in standard search and scanning down the results, requires little additional time and, based on this data, has a reasonable chance of finding a discount option for the identical product.
It’s also worth keeping in mind that the study represents a snapshot of a system that Google is actively evolving. AI Mode itself has seen meaningful shifts to its underlying models, its ranking factors, and its relationships with Search and Chrome, over the last year, and there’s no guarantee the patterns captured here will still exist six months from now. Productrise has itself indicated it intends to continue monitoring this discrepancy over time, building on both its July study and this one, meaning that shoppers, brands, and indeed anyone else should treat this as less of a conclusion and more of a beginning to a larger conversation.
A simple habit covers most of the practical risk here without asking you to become an expert on Google’s ranking systems. Simply speaking, before completing a purchase through AI mode, especially for a higher cost where a 20% disparity actually represents something in practice, take those additional few seconds to either click through to the alternate sellers panel, or reopen the same search in standard search. Given how often the two surfaces disagree on both seller and price, that quick cross-check has a real chance of finding you a better deal for the same product, and doesn’t require any more time than just accepting the first recommendation at face value.
A Note On What Comes Next
Studies like this tend to be opening salvos rather than closing arguments. Productrise has now seen fit to publish two data-driven examinations of AI Mode’s shopping behaviors within a few months of one another, first concerning the dearth of products presented by AI Mode versus standard search, and now about how those products are priced. With Google iterating on AI Mode so rapidly, including expansion of entry points on the search homepage, Chrome, and the search results page itself, it would not be surprising to see follow-ups to this study examining whether this price differential persists, narrows, or widens as the models and ranking algorithms change.
There are a few things the study and Google’s response make clear. Both AI Mode and standard Search rely on the Shopping Graph, and AI Mode shows far fewer products. Productrise’s data also shows that when the same products appear on both surfaces, AI Mode tends to show the higher price.
What is still unclear is why. It could be an intentional part of how AI Mode selects products, an unintended result of the ranking system, or a combination of factors. Until Google provides more detail, independent testing will remain one of the few ways to see how those choices are playing out.
Frequently Asked Questions
Q 1: What did the Productrise study actually find?
The study found that, when the same product appears in both Google’s standard search results and AI Mode, the AI Mode price is 21.6% higher on average. Further, the study found that AI Mode surfaces far fewer products overall and rarely surface the same products for identical searches.
Q 2: How large was the study, and how long did it last?
Productrise tracked over two million product listings across over one hundred thousand searches over 23 days, from August 9 to August 31, 2026, searching the same queries on the same day across the United States and United Kingdom in their research.
Q 3: Did Google respond to the findings?
Yes. Google informed Futurism it had not verified the accuracy of the claims made in the study, and that shopping results across its search products draw from the same underlying Shopping Graph data source, and that shoppers can click through any listing to compare prices across sellers.
Q 4: Does this mean Google is personalizing prices based on shopper data?
No, the study does not test or claim that. It measures what is presented in responses for identical queries at the same time. The study does not measure why such differences exist, nor whether particular shoppers see different prices based on their own data. Some commentary has associated the findings to the issue of surveillance pricing, but this is speculative, not something demonstrated by the findings.
Q 5: How often do AI Mode and Standard Search show the same products?
Very rarely. Only 1.28% of products appearing in the standard search carousel appeared in AI Mode for the same query on the same day.
Q 6: Is AI Mode always more expensive?
No. For matched products, prices agreed outright 61.9% of the time and disagreed on 38.1% of products, with AI Mode pricier in 68.4% of those disagreements and cheaper in the remaining 31.6%.
Q7: Why is the seller frequently different between the two surfaces?
Productrise found that the primary seller displayed for a matched product varied between the two surfaces 49.6% of the time. SEO commentators reviewing the study noted this implies that Google’s AI Mode considers factors beyond price and availability, including product feed quality and reviews, as well as third party citations, in selecting which seller to highlight.
Q 8: What can shoppers do to avoid overpaying in AI Mode?
Click through to a product listing to see the alternate sellers and prices presented in the knowledge panel, or cross-check a price by running the same search through standard search rather than relying on AI Mode’s top recommendation.
Q 9: What should retailers take away from this study?
A strong position in the standard product carousel does not guarantee the same visibility, seller placement, or price positioning in AI Mode. Brands should manually check how their products appear in AI Mode specifically, and should invest in richer product feed data, since this appears to influence AI Mode’s selections more than price does.
Q 10: Is this the first study on AI Mode’s shopping behavior?
No, Productrise published an earlier study in July, 2026, finding that AI Mode provides roughly 95% fewer product listings than standard search for the same queries. This new study builds on that earlier work with a new dataset focused specifically on pricing.
