What Happens If AI Starts Doing Most Online Searches: The Future of Google, Websites and Publishers

What Happens If AI Starts Doing Most Online Searches: The Future of Google, Websites and Publishers

Rating
—/5 0 ratings
Follow us on Google News
Perspectives10 min read

Internet search is entering a transformation comparable to the shift from web directories to Google. Rather than typing a few words and opening ten results, more and more users are asking complete questions and receiving a direct AI-generated summary. If this trend continues, Google, ChatGPT, Gemini and other systems will no longer be merely gateways to websites; they will become interfaces that read, compare and summarise the web on the user’s behalf.

For more on Google Search, see The internet after the internet: how much AI content is already online and what happens when machines write for other machines, which offers complementary details on this subject.

For publishers, the issue is substantial: if users receive an answer before clicking, organic traffic may fall even when a site’s information is being used behind the scenes to generate that answer.

AI is no longer a peripheral search feature

Google announced in 2026 that AI Overviews had more than 2.5 billion monthly active users, while AI Mode had surpassed one billion monthly users. The company says queries in AI Mode are, on average, about three times longer than conventional searches.

This points to a behavioural shift: people are no longer searching only for “restaurant in Constanța” or “DSG oil change”, but framing complex requests such as, “What is the best choice for my situation, and why?”

The traditional search-engine model was straightforward: the engine indexed the internet and directed users to relevant websites.

The AI model changes that relationship. The system can read several sources, compare them, extract the main ideas and generate an answer before the user opens a single site.

Data already indicate that AI reduces clicks

A Pew Research Center study of Google users’ browsing behaviour found that when an AI Overview appeared, users clicked on a conventional result in about 8% of visits. Where there was no AI summary, the click rate was nearly twice as high, at around 15%.

Moreover, only around 1% of visits to pages with an AI Overview were followed by a click on one of the sources cited directly in the summary.

These data do not prove that every site will lose traffic in the same way, but they point clearly in one direction: the more complete the answer on the search page, the fewer reasons users have to leave it.

The “zero-click” phenomenon could become the norm

A “zero-click search” is a search that ends without the user visiting an external site. Featured snippets, knowledge panels, maps and direct answers already existed before AI.

AI intensifies the phenomenon because it can answer more complex questions and combine several sources into a single result.

What this means for online publishers

For sites that depend on organic traffic, the shift could directly affect advertising, affiliate and subscription revenue.

  • fewer page views;
  • fewer advertising impressions;
  • fewer conversion opportunities;
  • less direct brand visibility;
  • greater dependence on the platforms that mediate access to information.

SEO is not disappearing, but it is changing

Conventional search-engine optimisation is not becoming irrelevant. On the contrary, page structure, domain authority, information quality, timeliness and clarity are becoming even more important, because AI systems need credible sources from which to extract answers.

The difference is that the goal is no longer only to “rank first”, but also to be one of the sources that AI selects and cites.

A new discipline is emerging: generative engine optimisation

Terms such as GEO — Generative Engine Optimization — and AEO — Answer Engine Optimization — are appearing increasingly often in the industry.

The idea is straightforward: content must be designed so that AI models can quickly identify answers, sources, data and context.

What kind of content is more likely to be used by AI

  • clear answers to specific questions;
  • original data and verifiable information;
  • a logical structure with subheadings;
  • tables and comparisons;
  • well-written FAQs;
  • explicitly cited sources;
  • regular updates;
  • genuine expertise and experience.

Why original sources are becoming more valuable

If ten websites rewrite the same news story, an AI system does not need to cite all ten. It will try to identify the most relevant, original or authoritative source.

This could heavily penalise publishers that rely almost exclusively on rewriting, while favouring material that brings its own information, explanations, comparisons, documents, interviews or analysis.

Brand becomes more important than the keyword

In a world where AI provides the answer directly, people will need reasons to look specifically for a website or publication.

Google has introduced and expanded Preferred Sources, a feature that lets users favour certain publications in Search and AI experiences. In 2026, the company announced that the number of users of the feature had grown more than sevenfold since the start of the year.

This is an important signal: the direct relationship between reader and brand may become as important as a position on a results page.

Google is trying to preserve the connection to the web

In October 2026, Google announced several changes to AI Mode and AI Overviews intended to make links to sources more visible: inline links higher up in the answer, dedicated sections for official sources, carousels for recent articles and more prominent links to related material.

This suggests that Google is aware of the tension between directly generated answers and the need to support the web ecosystem.

ChatGPT and other models are also changing the user journey

ChatGPT Search and similar systems can search the web and display answers with sources and citations. For users, the experience is closer to a conversation than to a traditional list of results.

That means that some searches which previously began on Google may start directly in an AI assistant.

AI traffic may be smaller, but sometimes more qualified

A user who clicks through from an AI answer may have a clearer intention. They have already read the summary and visit the site for detail, confirmation or additional information.

That means the number of visits may decline, while the value of each visit may rise in some niches.

Publishers will need to measure more than page views

In the new ecosystem, established indicators such as page views and average Google ranking will remain important, but they will no longer be sufficient.

  • AI mentions and citations;
  • direct traffic;
  • branded searches;
  • newsletter subscribers;
  • push notifications;
  • returning visitors;
  • conversions and time spent on the site.

Newsletters and push notifications become more important

If search-engine traffic becomes more volatile, publishers will seek to build direct relationships with their audiences.

Newsletters, push notifications, proprietary apps and subscriptions reduce dependence on platform algorithms.

Proprietary apps could become a strategic advantage

A publisher with its own app can maintain a direct relationship with readers through notifications, favourites, accounts and personalised content.

In an AI-dominated ecosystem, the direct channel becomes more valuable precisely because it does not depend on Google or a chatbot deciding to send traffic.

AI can also become a new distribution channel

Not all effects are negative. If an article is cited in AI answers, the publication may reach users who otherwise would not have discovered it.

For small and medium-sized brands, repeatedly appearing as a source can build authority even before the user visits the site.

The economic question: who pays for content?

If AI systems use articles to generate answers but send little traffic back, an economic problem emerges for publishers.

In 2026, Cloudflare introduced Pay Per Crawl, a system in beta through which website owners can charge AI crawlers for access. The concept also uses the HTTP 402 Payment Required response for requests made without accepting payment.

It is still too early to say whether this model will become standard, but it shows that the industry is trying to build an economy in which automated access to content can be monetised.

The web could divide between free, licensed and blocked access

In the future, websites may adopt different strategies:

  • allow AI crawlers free access in exchange for visibility;
  • allow only certain platforms;
  • charge for crawling;
  • license content directly to AI companies;
  • block automated access altogether.

What happens to websites that publish only generic content

These are probably the most vulnerable. If an article offers nothing different from dozens of other sources, AI can summarise the information without giving users a strong reason to visit the page.

Generic, mass-produced content with no original value may lose both traffic and relevance.

What kinds of websites could benefit

  • publishers with original information;
  • sites with their own data;
  • strong local sources;
  • experts and authors with a reputation;
  • sites with tools, calculators and interactive features;
  • brands with communities and a direct audience.

Local publishers have an advantage AI cannot invent

AI can summarise existing information, but it cannot independently attend a local meeting, take its own photograph at an event or call an official for a statement.

Original local content, investigations, documents and information obtained directly become more valuable precisely because they are difficult to replace.

What a publication should do in 2026–2027

  • write original articles, not just rewrites;
  • use clear sources and citations;
  • build authority in specific subjects;
  • optimise structure for clear answers;
  • use FAQs where they are useful;
  • build a newsletter, push-notification channel and proprietary app;
  • track traffic from ChatGPT, Gemini and other platforms;
  • measure branded searches and direct traffic;
  • publish material worth citing, not merely indexing.

What Google could look like in a few years

Google is unlikely simply to become “a chatbot”. The company is already trying to combine AI answers with links, carousels, traditional results, video, images and transactional features.

Search will probably become more conversational, multimodal and action-oriented, but the web will remain the infrastructure from which the system draws information.

Search becomes a conversation

In AI Mode, users can continue with follow-up questions without restarting their search. In 2026, Google also expanded Search Live, which enables interaction by voice and camera in more than 200 countries and territories.

This turns search from a sequence of independent queries into an ongoing dialogue.

The greatest risk for the web: losing the economic incentive

If websites no longer receive sufficient traffic and revenue, they may reduce investment in journalism, research and original production.

In the long term, that would also affect AI models, because they depend on a web rich in new and credible information.

Why AI platforms need a healthy web ecosystem

Google insists that it is developing features to help users reach original sources more easily. OpenAI displays citations and links in answers that use web search.

The central conflict of the coming years will be finding a balance between user convenience and the economic sustainability of information creators.

What this means for a site such as InfoBuz

For a digital publication, the strategy should not be to abandon SEO, but to expand it.

  • conventional SEO for Google;
  • clear structure for AI engines;
  • timely reporting and local data;
  • premium evergreen content;
  • a proprietary app and notifications;
  • a recognisable brand;
  • original sources and editorial transparency.

In a world with fewer clicks, authority and a direct relationship with the reader become strategic assets.

Conclusion: AI will not kill search, but it will change its economics

The future of search is not the disappearance of websites, but a change in their role. Websites will remain the source of information, but users may increasingly reach them less often directly for simple questions.

The publishers best placed to endure will be those that provide original information, build a brand, maintain direct channels to their audiences and become credible enough to be cited by AI engines.

Will Google Search disappear?

There is no indication that Google Search will disappear. More likely, it will become increasingly conversational and integrated with AI.

Do AI Overviews reduce traffic to websites?

Pew data show that users click less often when an AI summary appears, although the impact varies depending on the type of search and the site.

Is SEO still worthwhile?

Yes. SEO remains important, but it is expanding to include optimisation for answers and citations in AI systems.

How can a website be cited by AI?

By publishing original, clearly structured, verifiable and updated content, supported by sources and editorial authority.

What becomes more important than Google traffic?

Direct traffic, brand, newsletters, push notifications, proprietary apps and a community of readers.

Sources consulted

Etichete: AI SEOChatGPTChatGPT SearchGoogleGoogle SearchArtificial intelligenceonline publicationsorganic traffic
Share this article
COMUNITATEA INFOBUZ

Discuția cititorilor

Ai o opinie sau o întrebare despre acest articol? Spune-ne ce crezi.

0 comentarii

Fii primul care deschide discuția.

Lasă un comentariu

Email address este vizibilă doar administratorilor și nu va fi publicată. Câmpurile marcate cu * sunt obligatorii.

Te rugăm să păstrezi un limbaj civilizat și să discuți la subiect. Comentariile spam sunt filtrate automat.