Google’s AI-Generated Content Page Patent: How Search Could Change

Google has patented a system that could change what happens after a user clicks a search result.

The patent, US12536233B1, “AI-generated content page tailored to a specific user,” describes a system that can evaluate an organization’s existing landing page and, in some cases, give the searcher access to a new AI-generated page built around the user’s query and previous search activity. The patent was filed on January 3, 2025, granted on January 27, 2026, and is listed as active.

The important point is that Google is not merely describing an AI answer that appears above the organic results.

The patent describes a system that can sit between the search results and the website experience.

Google could identify a landing page, score it, decide that an AI-generated alternative should be offered, create a personalized page, and show a link to that page in Search.

That has major implications for SEO, ecommerce, landing-page optimization, conversion optimization and the future of organic traffic.

What does the Google patent actually claim?

The core claim is fairly specific.

Google claims a computing system that:

1. Receives a user’s search query.
2. Generates a search results page.
3. Identifies a result connected to an organization’s landing page.
4. Calculates a score for that landing page.
5. Compares the score against a threshold.
6. Generates an updated search results page when the threshold condition is met.
7. Adds a navigation link to an AI-generated page for the organization.
8. Displays the updated results to the user.

The key part is the landing page score.

Google is not claiming that every website will automatically receive an AI-generated replacement page. The claimed system first evaluates the existing landing page and uses the score as part of the decision process.

“calculating a landing page score for the first landing page”

Then comes the important step:

“generating an updated search result page based on the landing page score exceeding a threshold value”

In simple terms, Google’s patented system could work like this:

Search query → organic result → landing-page evaluation → score → threshold → AI-generated page option

That is the basic concept protected by the independent system claim.

Google could create a new page when the existing page is not good enough

The patent’s background explains the problem Google is trying to solve.

A search result can point to a landing page that is difficult to navigate. A user may struggle to find a product, use filters, understand the content or complete a purchase. Google says better relevance and easier navigation could improve usability and engagement.

The patent therefore moves beyond simply ranking the existing page.

Google describes the system as:

“automatically generating content pages that are specifically tailored to a user based on their search history.”

That sentence is important.

The proposed page would not simply be a copy of the company’s existing URL. The system could use the user’s current query, previous queries and other signals to create an experience designed for that particular searcher.

The page could use a user’s previous searches

One of the most interesting parts of the patent is personalization.

After a user chooses to view the AI-generated page, the system can retrieve contextual information from the user’s account. The patent specifically gives a previous query as an example of that context. Google can then combine the current query with that previous query and use the machine-learned model to generate the page.

The patent’s example makes the idea easier to understand.

A user might previously search:

• “best laptop for architecture”
• “best laptop for 3D modeling”
• a particular laptop brand

The next search could trigger a link to an AI-generated page customized around the user’s previous searches and interactions.

So imagine a user searches for a laptop.

The company’s normal landing page might show dozens of products.

Google’s AI-generated version could potentially understand that the user has already searched for architecture and 3D modeling laptops. The generated page could then emphasize products, information and actions that fit that search journey.

The patent describes the concept as a page customized to the user based on previous searches and user interactions.

What could the AI-generated page contain?

The patent gives several examples.

Product calls to action

An AI-generated page could contain a call-to-action leading to a product page belonging to the organization.

For an ecommerce site, that could create a path such as:

Search → AI-generated product experience → product page

The AI-generated page could act as a personalized shopping layer rather than simply sending every visitor to the same category page.

Product feeds

The patent also claims an embodiment where the generated page contains a product feed that gives an overview of products associated with the organization.

That could be useful when the existing landing page has a poor product-discovery experience.

AI chatbot

The generated page can also include an AI chatbot.

That creates an interesting combination.

A future Google-generated page could potentially contain:

personalized content + product information + product links + AI conversation

The user could arrive at a page that has already been adapted to the intent behind the search.

Dynamically generated page components

The patent goes further by describing dynamically generated content for individual components of the page.

“each component of the AI-generated page is annotated with dynamically generated content based on the user query.”

That suggests the system could modify different parts of the interface based on the query rather than producing one generic AI page for everyone.

What will determine the landing page score?

The patent lists several possible signals.

The landing page score can be based on:

• conversion rate
• bounce rate
• click-through rate
• page design quality
• content quality
• other landing-page performance metrics

That is one of the biggest SEO implications.

Google is not describing the score as a traditional ranking score alone.

It could evaluate the performance and quality of the landing-page experience.

The patent explicitly states that conversion rate can contribute to the score. It also lists bounce rate and click-through rate as possible factors.

It also allows qualitative factors such as:

“page design quality or content quality.”

The patent does not provide a public formula for the score.

It does not say that Google will assign a fixed number of points for bounce rate, another number for content quality and another number for conversion rate.

So SEOs should not treat the patent as a new confirmed Google ranking formula.

It describes possible system implementations.

What happens when the landing page crosses the threshold?

The threshold is central to the patent.

When the landing page score exceeds the threshold, the system can generate an updated Search results page with a link to an AI-generated page.

One example is particularly interesting for ecommerce.

The patent says the landing page score can exceed a threshold when the existing landing page does not have a product filter.

That creates a possible scenario:

A user searches for a product category.

Google finds a company’s landing page.

Google evaluates the page.

The page lacks useful product filtering.

Google determines that the page meets the relevant condition.

Google provides an AI-generated page option.

The AI page could then present products in a more useful format.

The patent does not say that Google currently applies such a system to every ecommerce search. It describes an implementation that Google has patented.

Google could use the page for future searches

The patent contains another important detail that can easily get overlooked.

The AI-generated page can potentially be used for future search.

That changes the idea from a one-time generated landing page into something potentially persistent.

A user could search for a product.

Google could generate a personalized page.

The user’s later searches could provide additional context.

The system could then use that context to produce another personalized experience.

The patent’s description also says the system can use past queries, current queries and other signals to curate search results and webpages.

That points toward a Search experience that remembers the user’s journey rather than treating every query as an isolated event.

Google could personalize the entire search experience

The patent is not limited to the AI-generated page itself.

It also describes an updated search results page containing:

• a personalized headline
• suggested filters
• suggested clusters

That suggests personalization could happen before the user even reaches the generated page.

For example, someone searching for running shoes might receive different suggested filters after Google understands prior searches.

One user might see:

Trail → Waterproof → Wide fit

Another might see:

Road → Lightweight → Men’s → Size 10

The patent does not prescribe those exact filters. They are an illustration of what the concept could enable.

The broader idea is that Search could adapt the interface itself around the user’s intent.

Google could use the technology across multiple organizations

Another important detail appears in the patent’s description.

Google says the system can generate landing pages for a plurality of content providers.

That suggests the technology is not limited to one specific company or website.

The system could potentially operate across many organizations whose pages appear in Search.

That creates a different model for organic search.

Today, Google generally sends a user to a page that the website owner created.

Under the patented concept, Google could potentially use the organization’s information and create a new interface tailored to the searcher.

The website remains the source of the underlying business information, but Google controls the presentation layer.

What does that mean for SEO?

This patent could push SEO beyond the traditional goal of ranking a URL.

The traditional model is simple:

Create page → optimize page → rank page → earn click → convert visitor

The patented model introduces another possible path:

Create page → Google evaluates page → Google generates personalized experience → user interacts with AI-generated page → user reaches product or conversion page

That does not make websites irrelevant.

In fact, the opposite could happen.

The quality of the underlying website could become increasingly important because Google needs content, products and other information from organizations to create useful generated experiences.

The patent explicitly says the system can use data from a data resource and an account profile to generate the AI-generated content page. The account profile can include user preferences, the current query and previous searches.

For SEO professionals, the implication is that content quality, structured product information, landing-page usability and conversion performance could all matter to the future search experience.

However, the patent alone does not establish that Google currently uses all of those signals as live ranking factors.

The patent could change what “ranking” means

The most interesting future scenario is not simply that Google generates another page.

It is that the winning webpage may no longer be the final experience the user sees.

A company could rank in Search because its page is relevant.

Google could then evaluate the page.

If the page experience is considered unsuitable for the particular query, Google could potentially create another interface based on the company’s information.

That creates two separate concepts:

Ranking the source

and

generating the experience

The first determines which organization or page enters the search ecosystem.

The second determines how the user interacts with that information.

That distinction could become important as Google adds more generative features to Search.

The patent is also designed around conversion

Google explicitly discusses improving content-provider performance.

The patent says the system can improve performance metrics such as total conversions and conversion rates.

That is significant because the system is not presented solely as a way to answer questions.

It is also presented as a way to improve the interaction between a searcher and a content provider.

An AI-generated page could guide a user toward a product, display a product feed, provide a chatbot and offer links to product details.

The result could look less like a traditional search result and more like a personalized digital storefront generated by Google.

The AI page could include previous-query information

The patent also allows the generated page to include information about a previous query and a sitelink to a product detail page.

Consider a user who first searches:

best laptop for architecture

Then:

best laptop for 3D modeling

Then:

MacBook Pro M-series

The next interaction could potentially use that history to create an experience that connects the user’s research journey.

Google’s own example uses a very similar sequence of laptop searches.

That is a significant shift from keyword-based personalization.

The system could understand the sequence of searches and use that sequence as context.

The underlying AI system can use multiple content types

The patent describes a machine-learned content generation pipeline that can include text, image, audio and video generators.

That opens the door to generated pages containing different media types.

A future page could therefore potentially combine:

text + images + products + video + audio + conversational elements

The patent does not say Google will use all of those elements on every page.

Rather, it describes a broader machine-learned generation system capable of handling multiple content types.

User feedback could improve the generated pages

Google also describes a feedback mechanism.

The system can receive feedback on generated assets and initiate the generation of updated or different assets.

The patent also discusses reinforcement learning and active learning using user feedback to help models learn user preferences and improve over time.

That suggests a potential feedback loop:

Generate → user interacts → collect feedback → improve model → generate better experience

If implemented at scale, the system could become increasingly adapted to how users interact with generated pages.

Could Google replace websites with AI pages?

The patent does not say that Google will replace websites.

That distinction matters.

The claimed system creates a link to an AI-generated page associated with an organization.

The AI page can also contain calls to action and links back to product pages.

So the organization’s website can remain part of the journey.

The bigger change is that Google could potentially control an additional layer between the search result and the organization’s traditional landing page.

That could give Google greater control over the user experience while still sending users to the company’s own product or detail pages when appropriate.

Is the patent already changing Google Search?

There is no basis in the patent itself to say that Google has fully deployed the claimed system across Search.

A patent describes technology that Google has sought to protect. It does not prove that every described feature has become a live Google Search feature.

That distinction is especially important for SEO analysis.

The patent was granted on January 27, 2026, and its status is listed as active.

But an active patent should not automatically be treated as evidence of a live ranking algorithm.

The right way to interpret it is as a technical signal of a direction Google has considered and protected.

What could the patent mean for websites in the future?

If Google implements the technology described in the patent, websites may need to think about their pages in a different way.

1. Landing-page quality could become more important

The patent gives Google several possible ways to evaluate landing pages, including conversion rate, bounce rate, click-through rate, content quality and design quality.

That could increase the value of a useful, well-designed landing page.

2. Poor UX could create an opening for Google’s AI layer

Google specifically describes generating an immersive page for queries that have poor targeting or a poor landing page.

A site that ranks but provides a weak experience could potentially become a candidate for an AI-generated alternative.

3. Search history could shape the experience

Google could use previous queries and other contextual signals when generating the page.

That makes the user’s search journey more important.

4. Product data could become even more valuable

The patent explicitly describes product feeds, product calls to action and product-detail sitelinks.

For ecommerce businesses, clean product information could become useful beyond traditional organic rankings.

5. SEO could move closer to experience optimization

The patent combines search, landing pages, machine learning, personalization and conversion.

That points toward a future where SEO is not only about getting a URL into Google’s results.

It could also involve making sure Google has enough high-quality information to build useful search experiences around the organization.

The biggest SEO takeaway

The most important part of US12536233B1 is not simply that Google can generate an AI page.

The bigger idea is the decision system around that page.

Google describes a process where Search can identify an organization’s landing page, calculate a score, compare that score with a threshold and potentially provide an AI-generated alternative.

That creates a possible future where Google does not always treat the publisher’s landing page as the final interface.

Instead, Google could decide:

“The organization is relevant to the query, but the current page is not the best experience for this user.”

The search engine could then generate a new experience using AI.

That is a much bigger idea than an AI summary.

It is a potential shift from ranking webpages toward generating personalized experiences around ranked organizations.

What should SEOs watch now?

SEOs should not start optimizing for a fictional “AI landing page score.”

Google has not published a formula for the score in this patent.

Instead, the patent gives a useful list of areas worth watching:

• landing-page content quality
• page design
• conversion performance
• bounce rate
• click-through behavior
• product filtering
• product information
• user intent
• previous search context
• personalized interfaces
• product feeds
• AI-assisted interactions

The patent itself makes clear that many of the described features are possible implementations rather than requirements for every version of the system.

That is why the patent should be treated as a directional signal, not a new Google ranking-factor announcement.

Final takeaway

Google’s AI-generated content page tailored to a specific user patent describes a future in which Search could do something very different from simply sending users to the best-matching webpage.

Google could identify an organization’s landing page, evaluate it, calculate a landing-page score and, when the relevant threshold condition is met, add a link to an AI-generated page.

That generated page could use the user’s current search, previous queries and other contextual signals.

It could contain product feeds, calls to action, product links, dynamically generated content and an AI chatbot.

The most important implication for SEO is simple:

Google may eventually have the ability to rank an organization’s webpage while generating a different, personalized interface for the user.

If that happens, winning organic visibility will no longer be the entire battle.

The quality of the underlying content, product information, landing-page experience and user journey could all influence how Google builds the next layer of the search experience.

And that could make the boundary between SEO, AEO, CRO and generative search optimization much thinner.