AI-generated landing pages? The end of traditional e-commerce? How Google's new AI patent is changing the rules of the game

Xpert Pre-Release

Online contact (Konrad Wolfenstein)

Language selection 📢

Published on: March 14, 2026 / Updated on: March 14, 2026 – Author: Konrad Wolfenstein

The end of traditional e-commerce? How Google's new AI patent is changing the rules of the game – Image: Xpert.Digital

Too high a bounce rate? How Google will soon replace faulty online shops with its own AI-powered pages

AI instead of a retailer's website: What e-commerce operators need to know about the new Google patent

To the general public, it sounds like just another highly technical document in the endless archives of the US Patent Office – but for the digital economy, it holds enormous potential for change. With patent US12536233B1, granted in January 2026, Google secures the legal foundation to fundamentally reshape online commerce. The core of the patent describes a mechanism by which the search engine evaluates retailer websites based on hard metrics such as conversion rate or the absence of product filters. If the assessment is negative, Google no longer redirects the user to the shop, but instead inserts its own, precisely tailored, AI-generated landing page.

Combined with the rapid developments surrounding AI and new, cross-platform commerce protocols, one thing is clear: Google is no longer content with simply being a guide on the web. The tech giant is poised to monopolize the entire purchasing process – from the initial search query to checkout – on its own platforms. What does this mean for online retailers, the future of search engine optimization, and antitrust law? And what strategies must merchants now employ to protect themselves against this erosion of Google's power?.

When the channel becomes the competitor: How Google is rewriting the rules of e-commerce with a single patent

A patent that is more than a patent

On January 27, 2026, the United States Patent and Trademark Office granted Google LLC patent US12536233B1, entitled "AI-generated content page tailored to a specific user." The European equivalent, EP4685671A1, was published concurrently and is still under review. To the general public, this sounds like just another technical patent in a sea of tens of thousands of Google patents. But such indifference would be a mistake. US12536233B1 doesn't describe an algorithmic optimization step, an improvement in speech recognition, or a new method for map display. It describes a concrete mechanism by which Google evaluates companies' landing pages and, in the case of a negative assessment, supplements or partially replaces them with its own AI-generated page. Combined with what Google is simultaneously rolling out in the market, this patent reveals itself as a legal cornerstone of a comprehensive strategy for controlling the entire digital purchasing process.

Anyone familiar with the history of the digital platform economy will immediately recognize the pattern. First, an open infrastructure is offered, upon which third parties build value. Then, as dependency grows, the platform itself begins to capture that same value. Amazon did it with its Marketplace and proprietary product categories. Apple did it with the App Store and its own system apps. Google has done it repeatedly with price comparisons, local search results, and travel bookings. Patent US12536233B1 is potentially the most consequential step in this long-standing logic of platform expansion.

What the patent essentially regulates

For patents, only the granted claim text counts, not the title, not the description, not the marketing language of the press releases. And the main claim (Claim 1) of US12536233B1 is remarkable in its precision:

Google receives a search query. Google generates a standard search results page with one result that links to an organization's landing page. Google calculates a "Landing Page Score" for this landing page. If this score exceeds a defined threshold, Google generates an updated results page. This updated page contains a navigation link to an AI-generated page for that organization. The updated results page is then presented to the user.

Only dependent claim 2 adds the personalization layer: contextual information from the user account, previous search queries, and processing by a machine-learned model. This means that the primary claim doesn't even require personalization. It already operates based on a poor landing page score plus a generated alternative page. This is a crucial legal distinction that is often blurred in public discourse. This patent does not apply to every AI search run, but rather to the precise combination of a quantitative quality assessment of a third-party page and the subsequent insertion of a Google-generated AI page into the search results.

The scoring system: Which signals does Google evaluate?

The most interesting and economically consequential passage of the patent describes the signals from which the landing page score is calculated. Google explicitly mentions: conversion rate (Claim 7), bounce rate (Claim 8), click-through rate (Claim 9), qualitative factors such as page design quality and content quality (Claim 10), general performance metrics (Claim 16), and, as a particularly concrete example, the finding that a landing page does not offer product filters (Claim 13).

This catalog of signals is remarkable for two reasons. First, these are all signals that Google already collects through its own infrastructure: Google Ads, Google Search Console, Merchant Center, Chrome data, and organic search. Google doesn't need to tap into any new data sources for this; it simply needs to consolidate existing data flows and apply a scoring mechanism. Second—and this is the real strategic challenge—these aren't classic SEO factors like text quality, authority, or thematic depth. They're about conversion optimization and interface design, precisely the categories in which many mid-sized retailers are structurally disadvantaged compared to large platforms and Google's own AI-powered pages.

The explicit example of missing product filters as a trigger for the score is no coincidence. It directly points to the e-commerce context: According to this logic, a retailer with a manageable product page who hasn't implemented a multi-level filter function is considered to be of insufficient quality, thus paving the way for Google's AI intervention. What can be marketed as a service from the user's perspective is, from the retailer's perspective, a circumvention of their own sales infrastructure.

The AI side: What Google is building instead of the retailer

The dependent claims and the patent description paint a concrete picture of the generated alternative page. It contains a CTA button to the retailer's product page (Claim 3), a product feed with an overview (Claim 4), an AI chatbot (Claim 5), dynamically annotated content based on the search query (Claim 6), personalized headlines and suggested filters (Claim 11), sitelinks to product detail pages (Claim 15), information about the user's previous search queries (Claim 14), and the explicit option to place the navigation link to this page in a sponsored content item (Claim 12).

The example in the patent is revealing: A user has previously searched for "best laptop for architecture" and "best laptop for 3D modeling." The AI page uses this search history to generate a tailored product overview—with the right filters, relevant products, and clear calls to action. The patent, therefore, does not describe a generic overview page. It describes a dynamically personalized purchasing interface distilled from the user's search behavior. And this page is inserted between the search results and the retailer's website.

The economic consequence is clear: The initial contact between potential buyer and product offering no longer takes place on the retailer's website, but on Google's own AI interface. Google's ability to monetize this touchpoint, including through sponsored placement (Claim 12), transforms the patent from a convenient technical feature into a business model.

Why this patent comes at the right time

Patent US12536233B1, viewed in isolation, might represent a medium-term opportunity that Google is legally securing. But it is not alone. It intersects with an economic ecosystem that Google has rapidly built over the past twelve months and integrates seamlessly into this infrastructure.

Google's AI Mode is well past the testing phase. With 75 million daily users worldwide, AI Mode is a mass phenomenon, not a laboratory experiment. Queries in AI Mode take two to three times longer than traditional searches because users ask follow-up questions, compare product options, and seek advice instead of simply clicking links. The search becomes a conversation that takes place entirely within Google's interface. Retailers who wait for users to click on their URL are waiting longer and longer.

In January 2026, at the National Retail Federation in New York, Google unveiled a wave of commerce products, all pointing in the same strategic direction. The Universal Commerce Protocol (UCP) is an open standard for AI-powered commerce, co-developed with Shopify, Etsy, Target, Walmart, Visa, and Mastercard. UCP-based checkout is already live: US users can purchase products from Etsy and Wayfair directly within AI Mode and the Gemini app without leaving Google. Shopify, Target, and Walmart are slated to follow. The Business Agent brings brand-specific AI chats directly into Google search results; Lowe's, Michaels, and Reebok have been live since January. Direct Offers allows advertisers to place exclusive discounts as Sponsored Deals within AI Mode, with pilot partners including PetCo, elf Cosmetics, Samsonite, and Shopify merchants. Since February 2026, shopping ad formats have been displayed directly in AI responses within AI Mode.

In addition, there's personal intelligence: Since January 2026, users in the US have been able to connect Gmail and Google Photos to AI Mode. Google explicitly showcases shopping scenarios as a use case: The system recognizes brand preferences from past purchases, reads travel destinations from Gmail booking confirmations, and suggests suitable products. Google VP Robby Stein publicly demonstrated how AI Mode recognizes a previous purchase history when searching for sneakers and uses this information to make new recommendations. This is precisely what Claim 2 of the patent technically describes: contextual information from the user account and previous search queries as input for AI page generation.

The structural logic behind the strategy

Anyone who puts the individual components together will recognize a coherent and very old logic of the platform economy: the systematic expansion into adjacent value chain stages. Google already controls search demand, the channel through which billions of purchasing decisions originate. The next logical step is to also own the purchasing experience itself.

Discovery, comparison, advice, offer, checkout: Google is building, step by step, an infrastructure that maps the entire customer funnel on its own interfaces. User Client Platform (UCP) standardizes data exchange between retailers and Google's AI agents. The Business Agent gives brands a voice in the conversational channel controlled by Google. Direct Offers transforms Google's AI interface into a trading platform. And patent US12536233B1 provides the legal safeguard that allows Google to continue operating even if a retailer hasn't sufficiently optimized their landing page for this new ecosystem.

The parallel to Amazon's strategy is striking. Amazon also initially built a sales infrastructure for third-party sellers, collecting data on demand and conversion performance, and then used this information to position and optimize its own products. Google uses the same structural logic, but operates on an even more fundamental level, because the search query itself takes place on Google's platform. Sellers on Amazon cannot do without Amazon. Sellers in the digital space can hardly do without Google – and it is precisely this dependency that creates the leverage.

The antitrust policy dimension

This development is taking place in a highly sensitive legal context. In August 2024, US Federal Judge Amit Mehta ruled that Google, with its search business, violated the rules of the Sherman Act and established an illegal monopoly in the online search market. Google challenged this ruling and formally filed an appeal in January 2026. The company argues that users choose Google out of conviction, not necessity, and that the ruling fails to consider the speed of innovation or the intense competition from AI startups and established players.

At the same time, Google faces further antitrust lawsuits. In January 2026, Judge Rita Lin ruled that a class-action lawsuit by consumers against Google's search monopoly would be allowed. According to the plaintiffs, Google systematically drove competitors out of the market through exclusive agreements with Apple, Android manufacturers, and telecommunications providers.

In this context, patent US12536233B1 appears as an interesting document from an antitrust perspective. A mechanism that allows Google to place AI-generated pages between merchants and buyers based on a self-defined quality assessment, and potentially market these pages as paid advertising formats, could be interpreted by regulators as a further instance of exploiting a dominant market position. The fact that Google has not yet faced any antitrust proceedings specifically in the commerce segment is partly due to the relative novelty of the relevant products. Should the AI Mode with UCP checkout and the landing page scoring described in this patent actually be rolled out on a large scale, this is likely to significantly increase antitrust scrutiny – particularly in the European Union, where regulatory practice is known to be more proactive.

🎯🎯🎯 Benefit from Xpert.Digital's extensive, five-fold expertise in one comprehensive service package | BD, R&D, XR, PR & Digital Visibility Optimization

Benefit from Xpert.Digital's extensive, five-fold expertise in a comprehensive service package | R&D, XR, PR & Digital Visibility Optimization - Image: Xpert.Digital

Xpert.Digital possesses in-depth knowledge across various industries. This allows us to develop tailored strategies precisely aligned with the requirements and challenges of your specific market segment. By continuously analyzing market trends and monitoring industry developments, we can act proactively and offer innovative solutions. The combination of experience and expertise generates added value and provides our clients with a decisive competitive advantage.

More information here:

Benefit from Xpert.Digital's 5 areas of expertise in one package – starting from just €500/month

Google's patent offensive: How an AI intermediary is challenging online commerce

What retailers and e-commerce operators need to understand now

The most important takeaway for retailers is initially a defensive one: The patent doesn't describe a direct attack on functioning, high-quality websites. The main claim refers to a link to an AI-generated page, not the disappearance of the retailer's own site. The more aggressive elements—sponsored placement, full product feed, AI chatbot—are contained in dependent claims. And not every granted patent is implemented. Google holds thousands of patents that never become products.

Nevertheless, it would be strategically unwise to leave this development at the patent level. What the patent describes is already becoming a reality, in part, through parallel product developments. UCP Checkout is live. The Business Agent is live. Direct Offers is running as a pilot project. Personal Intelligence is integrated into AI Mode. The technical and commercial infrastructure for AI intermediary pages is already operational. The patent merely provides the legal framework for a mechanism that is already manifesting itself in other forms.

For retailers, this means a shift in competition to dimensions that were previously of secondary importance. Landing page quality has primarily been a Google Ads concept, with the familiar Quality Score as its counterpart. The patent suggests that this logic could be extended to organic search results. If conversion rate, bounce rate, click-through rate (CTR), and design quality become criteria for whether Google includes an AI-generated page or not, then these are no longer just soft optimization recommendations, but hard survival signals.

Small and medium-sized retailers are particularly vulnerable, offering decent quality products but lacking the resources for continuous UX optimization, technical SEO, and top-tier Merchant Center data maintenance. These retailers would be systematically disadvantaged by Google's scoring logic, while their products would potentially remain visible on Google's own AI interface – now, however, within an interaction controlled and monetized by Google.

Product data as a new strategic resource

A structural shift, accelerated by the patent and the surrounding product ecosystem, concerns the relative importance of product data compared to product pages. In the classic SEO model, the product page was the primary value driver: it informed, persuaded, and converted. In the emerging model, Google can generate an AI-generated page based on Merchant Center data, even if the actual merchant website is of below-average quality. Those who maintain complete and accurate structured product data, prices, availability, compatible accessories, answers to frequently asked questions, and substitutes in Merchant Center increase the likelihood that their products will appear prominently on the AI-generated page, even if the user never visits the merchant's website itself.

This logic fundamentally changes investment priorities in digital commerce. Previously, the largest budgets went into website design, content production, and link building. In the emerging reality, structured product data, Merchant Center quality, and UX metrics are becoming equally important, and potentially even priority, areas for investment. A merchant with excellent product photos, precise attribute lists, and complete compatibility data, but a mediocre website, may rank higher on Google's AI interface than a merchant with a sleek website but poor data management.

GEO as a new discipline beyond classic SEO

The term Generative Engine Optimization (GEO) has become established as a complement to SEO, describing the optimization for AI-generated responses. Until now, the informational context has dominated: How does my brand appear in AI overviews? What content is cited in AI responses? Patent US12536233B1 makes it clear that GEO must necessarily also include the transactional and commercial context.

The crucial question shifts from "How do I rank in position 1?" to "How does my product appear in Google's AI-powered shopping interface?" This question is more complex because it cannot be answered by traditional SEO measures alone. It requires optimization based on conversion signals, complete data presence in the Merchant Center, visibility in price comparison and product feed structures, and participation in Google's commerce protocols like UCP. GEO in a transactional context is not a content problem. It's a data problem, a UX problem, and a partnership problem with Google.

The role of your own website in an intermediary world

The patent doesn't eliminate websites. The main claim describes a link to an AI-generated page, not the elimination of the destination itself. But it fundamentally changes the function of the website. In a scenario where the initial contact takes place on Google's AI interface, the website becomes the second step. If the user has already familiarized themselves with the product on the AI page, set filters, incorporated previous search queries, and clicked a call to action, they arrive at the retailer's site with significantly more specific expectations. Or they might purchase directly on Google's interface via UCP checkout and not even visit the website at all.

So what remains as the unique value of your own website? Exclusive content that Google cannot replicate: in-depth editorial advice, community features, loyalty programs, personalized offers for registered customers, after-sales services, and brand communication beyond the transaction. These are all dimensions that go beyond a product list and a checkout process, and which Google's AI-powered page cannot structurally offer. Merchants who invest in these dimensions of differentiation create a bulwark against intermediation.

At the same time, the possibility of sponsored placement (Claim 12) should be strategically considered. If Google actually markets the link to the AI-powered intermediate page in paid content, this opens up a new dimension for advertisers: the use of a Google-optimized, personalized landing page as an alternative to their own website, financed through the Google Ads system. Early pilot partners would, based on experience, have advantages in terms of the learning curve and placement, similar to the early adoption of Performance Max or other newer ad formats.

Industry-specific risk distribution

Not all market participants are affected equally. The patent clearly targets the transactional sector, specifically shopping search queries with a clear purchase intention. Purely informational content, B2B services without a product catalog, or highly specialized niche markets are less exposed, at least based on the current claim text.

Traditional e-commerce retailers, especially those with a broad, filterable product range, are particularly affected. The example of the missing product filters in the patent is too specific to be coincidental: it directly targets product listing pages (PLPs) in the fashion, consumer electronics, furniture, and household goods sectors. Google has already developed significant commercial interests in precisely these categories with platforms like Google Shopping, the Comparison Shopping Service, and now UCP Checkout. The patent complements these interests by providing a safeguard along the way.

For the German and European e-commerce markets, it's also worth noting that Google's UCP checkout rollout is initially focused on the US market. Etsy, Wayfair, Target, Walmart, and Shopify are US retailers. Their European counterparts, such as OTTO, Zalando, Douglas, and Saturn, are not among the partners mentioned so far. This gives European retailers a bit of a buffer, the length of which is unknown. Furthermore, given the regulatory sensitivity in the EU, where the Digital Markets Act already obligates Google to certain equal treatment measures, it's unclear whether and in what form the model described in this patent could be implemented analogously in Europe.

The interplay of innovation and power consolidation

A balanced economic analysis cannot consider patent US12536233B1 in isolation from Google's dominant position in the search market. A company that, according to a court ruling, holds an illegal monopoly in the search market patents a mechanism for evaluating the quality of third-party websites and, if necessary, replacing them with its own interface. This is structurally different from the same patent being held by a small startup or a marketplace operator without a search monopoly.

The argument that users benefit from improved UX on Google's AI-powered pages is undeniable. If a retailer operates a technically flawed page without product filters, with a high bounce rate and poor conversion, the primary victim is the user. Google's intervention could create genuine added value in these cases. However, the problem lies not in isolated incidents, but in the systemic shift in power. Google unilaterally defines the score threshold, Google decides which signals are accepted, Google designs the AI-powered page, and Google monetizes the touchpoint. The retailer has become a supplier of data and products, while Google controls customer access.

This logic is the core characteristic of platform economies: The platform operator gains value proportionally to the extent that competitors depend on its infrastructure. Google has built this position in the search market over decades. Patent US12536233B1 is the most direct expression to date of the ambition to extend this position into the e-commerce purchasing process.

Strategic options for action: A framework for companies

In light of this analysis, some strategic conclusions can be drawn for companies in e-commerce and related areas, without resorting to knee-jerk reactions or panic optimization.

First, landing page quality should be understood as a strategic investment, not merely a technical hygiene issue. The signals mentioned in the patent are measurable and optimizable. Conversion rate, bounce rate, click-through rate (CTR), design quality, and filter options are not abstract metrics, but concrete areas for action. Companies that seriously monitor these metrics reduce the risk of receiving a negative landing page score, regardless of whether Google implements the patent in production or not.

Secondly, product data in the Merchant Center is gaining a new dimension of strategic relevance. Complete, up-to-date, attribute-rich, and accurate product data is no longer just a prerequisite for Google Shopping ads. It is the raw material from which Google's AI-generated page is potentially created. Those who work diligently here will appear better on the AI interface, even if the user never visits their website.

Third, companies should systematically expand GEO to include transactional and commercial queries. Analyzing which shopping search queries already trigger AI-powered fashion results with product recommendations is a key task for the current quarter. Those who understand these queries and how their brand is represented within them can optimize accordingly.

Fourth, companies should identify and strengthen the unique value layers of their own website. In-depth consultation, community, loyalty programs, after-sales excellence, brand experience, exclusive content: these are the dimensions that an AI-powered site cannot structurally replicate and which therefore establish the long-term value of an independent web presence.

Assessment: What this patent is and what it is not

US12536233B1 is a granted patent with specific claim text that protects a specific mechanism, not a letter of intent or a marketing document. Not every granted patent is implemented. Google's patent portfolio includes thousands of documents that never became products.

At the same time, it's not just any patent appearing at any random moment. It emerges at a time when Google is rolling out precisely the infrastructure that the patent technically describes: UCP Checkout, Personal Intelligence, Business Agent, and Direct Offers. The correlation between the patent and product development is far too precise for a product to have appeared by chance.

The patent is therefore best understood as legal protection for a commercial strategy that is already underway. Google's goal of integrating the entire purchasing process onto its own AI-powered interfaces is documented by the NRF announcements of January 2026 and the parallel product launch. The patent protects a portion of this infrastructure against imitation and provides Google with a legal basis for future product developments in precisely this area.

In practical terms, this means that anyone who views the patent as an isolated document and hopes Google never uses it in production is missing the point. However, anyone who understands that the patent is both a symptom and an integral part of a broader strategic movement will have a clearer understanding of what to expect. Not the obsolescence of their own website, but certainly a world where the user's direct path from the search bar to the merchant's page is increasingly enhanced with Google's own AI layer – and where the quality of their presence in the Merchant Center and on the landing page determines whether Google redirects the user or retains them.

The race for the customer journey in digital commerce has begun a new round. And Google – which has always been its structural advantage – is controlling the starting signal.