The Maze: Google is making website links more prominent inside its AI-generated search answers. The October 15 update affects publishers, retailers and creators competing for visits from Search. Some changes are rolling out; others remain tests. Google has not named every new market or shown that the redesign increases website traffic.
Source: Google. Google demonstration of refund-form and refund-rule links above an AI answer, illustrating a planned official-site section test.
The Details:
Links move closer to the answer. Google is putting embedded links higher in responses more frequently. These are clickable words within an answer, so readers can open a relevant website while reading. Links at the end of some answers also gain larger, more visual treatment with thumbnails. The distinction matters for a business publishing product information or advice: being cited creates a possible route to its website, but the reader still has to choose to leave Google. Neither placement nor a larger image establishes that the business received a visit.
Different links lead to different places. A planned test would put official destinations for completing a task in a dedicated section near the top of an answer. Google's illustration shows links to an airline refund form and refund rules before the explanation below. It is a demonstration, not a verified airline service. A separate test makes creator and publisher names clickable on video results in AI Mode, Google's conversational search experience. Those names lead to Google-hosted Search profiles that gather a creator's work. They do not take the reader straight to the creator's own website. Google has not disclosed the official-site selection criteria or a country list for these tests.
The article cards are an expansion. Google is extending rows of timely article links to additional locations and languages, without listing them. These cards can highlight Preferred Sources: websites a reader has chosen as favorites in Search settings. Google introduced those labels and article carousels in May. October therefore expands an existing discovery route. Google also says the number of Preferred Sources users is more than seven times its start-of-year level. It supplies neither the starting count nor a precise definition of use. That growth figure measures adoption of a feature, not extra visits to a publisher or orders for a retailer.
MarketMaze Take: The useful change is a clearer route from an AI answer to another page. The unresolved business question is whether readers take it. For businesses in any supported market, the relevant sequence remains simple: a link appears, someone opens the destination, and that visit may produce an order or another valuable action.
Check the destination before valuing the exposure. The official-task section could shorten a reader's journey to a business-controlled page. The creator-name test instead sends the reader to another Google page first. Both make a name easier to find, but they offer different routes to an owned audience. That difference is more useful than treating every new clickable surface as equivalent traffic.
Keep the measurement claim narrow. Earlier links could send more visitors, and that is the strongest case for the redesign. But the available reporting contains no measured referral uplift from these changes. A sevenfold increase in feature users cannot fill that gap. Until completed visits and their outcomes are measured, a better-looking link is an opportunity, not evidence of recovered sales.
