Artificial intelligence has rapidly changed how people find information online.
A search can now produce a traditional list of results, an AI-generated summary, a conversational answer, or a combination of information from multiple sources within seconds.
As this transformation accelerated, predictions that SEO was becoming obsolete became common. New terms such as GEO, or Generative Engine Optimization, and AEO, or Answer Engine Optimization, also gained significant attention.
The reality is less dramatic and considerably more useful.
SEO is still relevant. What has changed is the environment in which discovery happens.
Google itself states that generative experiences such as AI Overviews and AI Mode remain grounded in its core search and quality systems to retrieve relevant information.
Crawling, indexing, useful content, authority, site architecture, and user experience therefore continue to matter.
However, optimizing only to achieve a position within a list of links no longer represents the entire strategy.
SEO did not disappear because of artificial intelligence
For many years, SEO was often reduced to the idea of ranking a page among the first results for a particular keyword.
That model never fully represented how search engines worked, and artificial intelligence has made it even less accurate.
New search experiences can interpret complex questions, identify different aspects of intent, and retrieve information from multiple sources before producing a response.
Google's AI Overviews and AI Mode, for example, can use a technique known as query fan-out.
Instead of processing only the original query, the system can run several related searches to gather complementary information.
A search about selecting a CRM platform for a small business might lead to parallel searches involving pricing, integrations, implementation complexity, automation capabilities, and suitability for different team sizes.
This creates opportunities for a page to be discovered even when it does not literally match the original query.
The implication is significant: SEO depends even less on exact keyword repetition and even more on answering a real need effectively.
What has not changed in SEO
Before discussing artificial intelligence, it is important to distinguish evolution from disruption.
Search systems still need to find, access, understand, and index a page before they can present it to users.
That means several fundamentals remain important:
Crawling and indexing: search systems must be able to access the important pages of a website.
Information architecture: content needs to be organized and connected through understandable internal links.
Technical quality: performance, mobile usability, stability, and accessibility remain part of a strong user experience.
Helpful and reliable content: pages need to satisfy user intent and provide information that genuinely helps.
Authority and experience: demonstrating real knowledge of a subject continues to matter, particularly in competitive industries.
Structured data: it can still help search engines understand specific information and enable eligible rich results.
What does not exist is a new secret code for "AI SEO."
Google states that websites do not need a special AI schema, an artificial writing style designed for generative models, or files such as `llms.txt` to appear in Google's generative search experiences.
The foundation is still SEO.
Generic content has become even less valuable
If one change deserves particular attention, it is the economics of content.
For years, businesses could publish articles that mainly reorganized information already available across dozens of other websites.
Articles such as "10 ways to improve your marketing," "5 reasons to invest in SEO," or "what is digital marketing?" could generate traffic simply by covering a keyword reasonably well.
Artificial intelligence can produce this type of summary within seconds.
That makes commodity content increasingly difficult to justify.
Google now uses this concept in its own guidance for generative search. Non-commodity content is more useful because it provides something that cannot easily be reproduced from the same information already available everywhere else.
In practice, this increases the value of first-hand experience, proprietary analysis, internal data, real comparisons, case studies, methodologies, tests, practical examples, and well-supported viewpoints.
A page that merely summarizes what other pages have already said provides little reason to become a reference.
A page that adds original knowledge serves a much clearer purpose.
Keywords still matter, but their role has changed
Artificial intelligence does not mean abandoning keyword research.
Keywords still help companies understand how people search for products, services, problems, and information.
The difference is that optimizing a page should not mean repeating a specific phrase a predetermined number of times.
Modern systems can understand synonyms, context, relationships between concepts, and different ways of expressing the same intent.
Keyword research therefore increasingly functions as demand and intent research, rather than a list of terms that must appear repeatedly on a page.
An effective strategy begins by identifying the user's central question and the information necessary to answer it.
The content can then develop the subject naturally.
Search is becoming more conversational
User behavior is changing as well.
In traditional search engines, someone might reduce a business problem to a short query such as:
"Google Ads agency Curitiba"
In a conversational interface, that same person might ask:
"I run a B2B company in Curitiba, spend around R$5,000 per month on paid media, and need to improve lead generation. What should I evaluate before hiring a Google Ads agency?"
The second query contains far more context.
This benefits content that explains problems comprehensively and demonstrates expertise in specific situations.
Service pages remain important, but educational content gains another role: helping search and AI systems find information that can support different stages of a decision.
Clear structure has become even more important
There is no need to write for robots.
However, organized writing helps both users and automated systems understand information.
Descriptive headings, coherent subheadings, focused paragraphs, and clear relationships between topics improve content comprehension.
That does not mean artificially splitting every answer into tiny pieces for AI.
Google explicitly states that there is no requirement to "chunk" content to appear in generative search.
Structure should exist because it improves usability.
A useful principle remains straightforward: each section should clearly answer a relevant part of the subject.
Being crawlable now involves more than Googlebot
This is one of the most practical technical changes.
For many years, Googlebot and Bingbot were the primary crawlers considered in SEO discussions.
As AI tools have expanded, additional crawlers have become part of the discovery ecosystem.
ChatGPT Search, for example, uses OAI-SearchBot to discover pages that may appear in search results and AI answers containing links.
This crawler is separate from GPTBot, which is associated with crawling content that may be used to help train OpenAI's generative AI foundation models.
A company can therefore allow OAI-SearchBot to support visibility in ChatGPT Search while making a separate decision regarding GPTBot.
For websites using Cloudflare, firewalls, or other bot-protection systems, robots.txt is not the only consideration. Security rules can also prevent these crawlers from accessing the site.
Crawler management has therefore become another component of digital visibility infrastructure.
Structured data is still useful, but it is not an AI shortcut
Structured data remains useful because it helps search engines interpret specific information about pages, products, businesses, articles, events, and other entities.
But there is no special schema that automatically causes a page to be selected by an AI system.
Adding more markup also does not guarantee greater presence in generative answers.
The principle remains unchanged: use structured data when there is a real application and ensure that the markup accurately represents the visible content of the page.
Authority needs to be demonstrable
Another consequence of abundant AI-generated content is that publishing large volumes of material is no longer an obvious competitive advantage.
If almost any company can generate dozens of articles quickly, volume alone provides little differentiation.
This increases the value of demonstrating who is responsible for the content and why that source should be trusted.
Identified authors, comprehensive company pages, demonstrable expertise, case studies, original research, clients, methodologies, and reliable sources all help establish credibility.
There is no official formula stating how many brand mentions are required for an AI system to recommend a company.
Artificial mentions, content produced purely for volume, or fabricated references do not create a sustainable strategy.
Authority is still built through knowledge, reputation, and consistent presence.
SEO measurement is changing as well
For many years, SEO measurement focused primarily on rankings, impressions, clicks, and organic traffic.
Those metrics remain important.
Generative search, however, adds another dimension.
In 2026, Google introduced dedicated Search Console reporting for visibility within generative features, allowing website owners to analyze information such as impressions, pages displayed, countries, devices, and performance over time.
This reflects a broader change in the definition of visibility.
A company can now participate in a user's research journey even when that user is not interacting with a traditional blue-link result.
SEO evaluation therefore needs to combine visibility, traffic, and, most importantly, business outcomes.
An increase in sessions without meaningful results may provide less value than a smaller amount of highly qualified traffic.
Can artificial intelligence create SEO content?
Yes.
The problem is not the tool used to produce the content.
The issue is why the content was created and whether the final result provides genuine value.
Google does not prohibit content simply because artificial intelligence helped produce it.
However, using automation to publish large volumes of low-value pages primarily to manipulate search results can violate its policies against scaled content abuse.
AI can assist with research, organization, editing, comparison, and production.
Expertise, editorial judgment, and responsibility for the published information still require human oversight.
Producing more content has become easy.
Producing something worth finding remains difficult.
What about GEO and AEO?
Terms such as Generative Engine Optimization and Answer Engine Optimization can be useful when discussing a particular area of search strategy.
They highlight the fact that companies now need to consider how their information appears within AI-generated answers.
Problems arise when these terms are presented as replacements for SEO or as collections of secret techniques designed to manipulate generative models.
For Google, the company itself states that optimizing for generative search remains part of SEO.
Other platforms may have technical differences, including specific crawler policies, but the underlying principles remain similar: accessible, understandable, reliable, and relevant information must exist before a system can recommend it.
So, what has actually changed?
The biggest change is not that businesses should abandon SEO.
The objective has expanded.
SEO is no longer only a competition for positions on a search results page. It is increasingly part of an ecosystem in which search engines, generative answers, and AI assistants can find, interpret, combine, and recommend information.
In this environment, several priorities become even clearer:
create knowledge that adds something useful to the web;
build websites that can be easily crawled and understood;
demonstrate expertise and authority in verifiable ways;
monitor how the brand appears across different discovery environments;
and connect organic visibility to real business outcomes.
Technical SEO is still necessary.
Keyword research is still necessary.
Content is still necessary.
Links, architecture, structured data, and user experience are still necessary.
Artificial intelligence has not eliminated the fundamentals of SEO. It has made superficial SEO considerably less effective.
Companies that understand this distinction do not need to choose between optimizing for Google and optimizing for artificial intelligence.
They need to build a digital presence strong enough to be found, understood, and considered relevant by both.
