Google adds more source links

Google is trying to answer the biggest complaint about AI search — that it keeps people inside the box and starves the open web of clicks. This week, the company rolled out more visible source links inside AI Overviews and AI Mode, with new ways to jump to publishers, brands, forums, and expert sources. That matters because AI search has been stuck in an awkward middle state. Users like fast summaries, but publishers, retailers, and creators want proof that these summaries still send traffic outward. Google’s answer is basically: make the links harder to miss. ### What exactly changed? The new design pushes more outbound links into the experience itself. (blog.google) Google highlighted inline links inside AI-generated text, “Further exploration” panels, richer source callouts, and more obvious paths from AI Overviews into AI Mode. In Chrome desktop, AI Mode also opens clicked pages side-by-side so users can read a source without losing the conversation. ### Why is Google doing this now? Because the trust problem is real. AI answers are only useful if people can check the underlying sources, compare viewpoints, and keep browsing the web instead of treating the summary like a final authority. Google’s own product language leans hard on that idea now — “explore the web,” “connect with authentic voices,” and keep context while visiting sites. (blog.google) ### Does this fix the publisher problem? Not by itself. More visible links can improve the odds of a click, but they do not guarantee that cited sites get meaningful traffic. The bigger issue is distribution: AI Overviews already appear at scale, and Google has said they’re used by more than 1 billion people, so even small interface changes can matter — but only if users actually leave the summary. (blog.google) ### So should sites just add more schema? Turns out, probably not. Ahrefs tracked 1,885 pages that added JSON-LD schema between August 2025 and March 2026, matched them against 4,000 controls, and saw no major citation lift across Google AI Overviews, AI Mode, or ChatGPT. That is the cleanest recent evidence against the idea that markup alone wins AI citations. ### Why didn’t schema move the needle? (blog.google) Because schema helps machines understand a page, but it does not make the page worth citing. AI systems still need content that looks authoritative, relevant, and easy to extract. Ahrefs’ earlier work points the same way: 86% of AI Overview citations came from pages already ranking somewhere in Google’s top 100, and the median cited URL ranked No. 3. Basically, classic search strength still carries a lot of weight. (ahrefs.com) ### What does that mean for ecommerce and beauty brands? It means the boring stuff still matters most. Product pages need clear comparisons, strong category authority, original information, and signals that other sites trust them. Technical cleanup is still worth doing — broken schema is not a flex — but it looks more like table stakes than a shortcut into AI answers. That is especially relevant for shopping-heavy queries, where Google is trying to keep AI Mode useful while still sending people to merchant and review pages. (ahrefs.com) ### What should marketers watch next? Watch click behavior, not just citation counts. Google is redesigning the surface so sources are more visible, but the real test is whether those links earn visits, sales, and repeat discovery. If they do, this is a meaningful concession to the web. If they do not, then “more links” is mostly cosmetic. (blog.google) ### Bottom line? Google is admitting that AI search needs to show its work. But the early data says winning those citations still depends less on clever markup and more on being the page Google already trusts. (blog.google)