New displays and tests may create referral opportunities, but the announcement does not establish a traffic lift.
PPC Land reports that in a post dated 15 October, “More ways to connect with the web from Search,” Google’s Hema Budaraju announced five changes to AI Mode and AI Overviews: inline links shown higher in responses more frequently; a test of a dedicated section at the top of an answer for “official” sites, with context on why each link is useful; news carousels for timely articles rolling out to more locations and languages; larger end-of-response links with image thumbnails; and a test of clickable creator and publisher names on video results in AI Mode that link to their Search profiles. Google said the number of people using Preferred Sources has grown by more than 7x since the start of the year, with no baseline.
PPC Land stresses what is missing: no numbers on how many responses carry inline links, no test sizes or markets, no criteria for who counts as “official,” and nothing on whether the changes send more visits. Two of the five items are explicitly tests.
In plain English
“AI Overviews” and “AI Mode” are Google’s generated answers in Search. An “inline link” is a link inside the answer text. “Preferred Sources” lets people choose outlets they want to see more of.
Why it matters
Google is adding more ways for AI answers to send people out, but gives no traffic data. Publishers and brands should measure referrals from these surfaces themselves before changing strategy.
More visible links create opportunities for referrals, but they do not measure whether people take them. An answer can offer additional links while still satisfying a user's immediate need on the search page. The business effect depends on actual visits and what those visitors do afterward.
For a publisher, the useful comparison is therefore a consistent set of traffic and engagement measures over time. Separate changes in impressions, referrals and downstream actions where the available analytics allow it. A rise in the number of people using a search feature does not directly establish a rise in visits to an individual publication.
The reported Preferred Sources growth also lacks a starting total. A multiple describes relative growth, not the absolute size of the audience or its importance to a given website. Treat it as a signal of product adoption rather than as a traffic forecast.
Rollout and test labels should remain visible in planning. An experiment with official-site links is not a guarantee that every organization will receive the same treatment. Likewise, an expansion to additional languages or locations does not mean universal availability. The announcement gives publishers concrete surfaces to watch, while leaving the central commercial question open: whether the changes produce useful, sustained audience visits.