Google AI Search is making a conspicuous adjustment to the way its machine-written answers display the web. Rather than leaving websites tucked away as supporting material, Google says it is placing more links within the answer itself, testing special shortcuts to official sites and widening the visibility of news links. These are changes to the presentation of its AI search experience, not a declaration that traditional search results are returning to their old format.
The announcement, made on 15 October 2026, matters because an AI answer can satisfy a question before someone visits the publisher that supplied the information. In its official product update, Google acknowledges the importance of exploration beyond the generated response. It says it has increased inline links, begun testing more prominent direct links for specific tasks and expanded article carousels. Those claims describe features and tests, not audited evidence that publishers will gain readers.
Google AI Search is changing where the source appears
The central difference is one of timing. A person reading a conventional search page generally chooses a website before receiving a detailed explanation. In an AI-generated answer, the explanation appears first, so the reader may never need to investigate the source. Putting a relevant link directly beside the sentence that uses the information gives the reader a more natural opportunity to verify, disagree or continue. That does not guarantee that the reader will follow it, but it reduces the distance between an assertion and the material behind it.
Google says it will show some official destinations in a dedicated section at the top of an AI response. Its example is the situation in which a person wants the correct site to complete a specific task. That distinction matters: an answer about passport eligibility and the actual government page where an application starts are not interchangeable. For practical searches, the destination may be more valuable than an eloquent summary of the procedure.
The company is also expanding prominent article carousels for timely topics into additional countries and languages. The feature may expose readers to a range of publishers without requiring them to abandon the AI experience first. Google has mentioned richer links to related articles and experiments with clickable profiles for video creators. These are separate interface changes with different audiences, not a single new ranking formula.
The tension is familiar to publishers. LiveAIWire has previously examined the pressure that AI Overviews place on news organisations. If a generated answer answers everything, a publication may win a citation but lose the visit that supports its business. A source appearing visibly is better than a hidden attribution, yet the commercial consequences depend on what actual readers do.
What website owners can and cannot assume
Google’s Search Central documentation says AI Overviews and AI Mode draw on the same underlying search systems and that eligible pages generally need to be indexed and suitable for snippets. There is no special technical requirement to appear as a supporting link in these experiences. That is an important corrective to the idea that publishers must add a secret AI-specific file or buy a new form of optimisation to qualify.
The documentation also says conventional search best practices still apply: accessible pages, useful written content, internal linking, sensible page experience and structured data that matches what people actually see. None of those elements promises inclusion. An indexed page can still be ignored for a particular answer, and an AI response may draw together material from several searches rather than replicate a list of results.
For site owners, the natural question is whether a more visible citation becomes a meaningful visit. Google has described the opportunity, but its October announcement does not provide independent before-and-after click data for publishers. It would therefore be premature to claim that this update reverses the loss of visits associated with AI-led search experiences or fixes the economics of reporting.
A useful measurement plan would separate impressions, conventional clicks, referrals associated with AI features where measurable and subsequent engagement on the site. It would also distinguish between a broad explainer that a chatbot can summarise and an original investigation, interactive tool or local service that has value beyond a short answer. The difference concerns user behaviour, not just a search-engine setting.
A clearer link is still not the same as a reliable answer
There is a second reason visible links matter: readers need a route back to the evidence. A link that appears beside a factual sentence allows someone to check whether a source really supports the claim. That is particularly important when a generated answer combines statistics, advice and interpretation. Good placement cannot itself prevent an AI mistake, but it makes checking less awkward and potentially more likely.
Readers should still examine whether a cited page actually says what the answer suggests. The presence of an official-looking source does not certify every sentence surrounding the citation, just as a tidy graphic does not settle a disputed claim. Official-site shortcuts are most useful when they deliver the destination the user asked for rather than another machine-generated explanation of it.
The wider test for Google will arrive in ordinary browsing: whether people can find original reporting, practical tools and authoritative pages without fighting the AI summary for attention. The issue sits at the heart of the debate about whether publishers can survive a zero-click web. Better links offer a plausible improvement to navigation. Whether they restore meaningful attention to the sites that produce the information is a question for evidence, not slogans.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity and the social impact of emerging technology. LiveAIWire is an independent, human-led technology publication using AI-assisted research, editorial production and original AI-assisted editorial illustrations under his direction.