What Google Covered at Search Central Live Milan
Google held its Search Central Live event in Milan on June 18, and the agenda was packed. Chunking, site-wide quality, paywalls, subscriptions, the value of an AI Overview click, vibe coding, a couple of new Search Console features. On paper it looks like a scattershot of unrelated topics. Listen to them together, though, and the same idea keeps surfacing. There’s no separate game for AI search. The work that earned visibility before is the work that earns it now, and Google spent a day saying so from different angles.
No separate playbook for AI search
The clearest version of that message came up around content. Google drew a hard line between commodity and non-commodity material. Commodity content is the generic stuff that already exists in near-identical form on a thousand other sites, the basic definitions and broad how-to pages, and Google has no reason to cite any one source for it. Non-commodity content brings something that isn’t already everywhere, like original analysis, first-hand experience, or proprietary data, which is what gives an AI answer a reason to point at you. The team also flagged a restrictive stance on synthetic, programmatic text produced at scale, which it files under scaled content abuse. The fuller version of that argument is in our piece on why authentic coverage wins in AI search.
Chunking got a lot of attention too. Before Google feeds your page into an AI system, it breaks the content into discrete segments, and whether a given paragraph can stand on its own decides whether it gets pulled into an answer or skipped over. Dense, undifferentiated walls of text are hard to chunk cleanly. Clear headings and self-contained paragraphs are easy. It’s the same structural advice that always helped with featured snippets, now load-bearing for AI extraction. We got into which formatting habits actually help, and which are myths, in our breakdown of the AI SEO hacks Google says don’t work.
The clicks debate, and a new way to settle it
One of the most-discussed slides was about what happens to your traffic when an AI Overview appears. Google’s position is that the clicks you still get are higher quality, meaning people who do click through tend to spend more time on the site. The catch, which the room noticed, is that the slide came with no numbers at all. Independent studies tell a tougher story on volume, with click-through on the top result dropping sharply on overview queries. The two views can coexist, and we walked through how in our look at the real value of an AI Overview click.
The more useful news for measuring any of this was a Search Console update. Google is rolling out AI reporting that isolates impressions and clicks from AI Overviews and AI Mode, along with a setting to include or exclude your site from AI features. For the first time you can see your own AI numbers rather than argue from someone else’s sample. What the report shows and how to use it is covered in our post on how Google now reports your AI search visibility.
Two features that change who gets seen
Milan also brought two features that affect which sites surface in the first place. The first is Preferred Sources, which lets readers nominate the publications they want to see more of in Google’s results. It started in Top Stories and has been expanding into AI Overviews and AI Mode, and it puts a thumb on the scale toward brands people already know and seek out by name. How it works and how to get listed is in our piece on how readers can pick the sites Google’s AI favors.
The second is subscription linking for publishers, which surfaces a reader’s existing subscriptions inside results through Google’s Reader Revenue Manager. It’s narrower than it first sounds, since it helps with subscribers you already have rather than winning new ones, and it doesn’t lift your rankings. We separated what it actually does from what people assume in our breakdown of what Google’s subscription linking really does.
A reality check on vibe coding
Google also closed a loop on something that’s been hyped all year, which is vibe coding, the practice of building tools and scripts by prompting an AI instead of writing the code yourself. The guidance was blunt. AI can knock out a basic script, but someone still needs to understand what it produced, because the long-term costs land in security, maintenance, and architecture, not in the first five minutes of writing. Google steered people toward its official endpoints, like the Search Console API, over improvised scrapers. The full argument, including where this bites SEO teams specifically, is in our post on why vibe coding still needs someone who can code.
What the day came down to
Put it all back together and Milan splits into two buckets. The advice about content and clicks was Google repeating itself, on purpose, because the fundamentals haven’t moved. Quality, original work, and earned authority still decide whether you show up, in AI answers as much as in blue links. The new material was tooling and features, the AI reporting, the AI settings, Preferred Sources, and subscription linking, which change the mechanics around the edges without changing what gets rewarded at the center.
For anyone deciding where to spend effort after an event like this, the answer didn’t change. Build the kind of authority that gets you cited and referenced, then use the new Search Console reporting to watch it show up. The features will keep arriving. The thing they all point back to is the same work that good link building and original content have always done.