For fifteen years, Search Console answered exactly one question: how is my website doing in Google? That is no longer what it is. In the space of eight months Google has bolted on a report for your visibility inside AI Overviews and AI Mode, and a whole new property type that tracks how your Instagram, TikTok, X and YouTube posts perform in Search — a property you can verify without owning a website at all. Most teams haven’t opened either. Here is what actually shipped, what the data does and doesn’t tell you, and what it should change in your monthly reporting.
SEO · Measurement
What Actually Changed in Search Console This Year?
Two things matter, and they arrived two months apart. On 3 June 2026, Google introduced Search Generative AI performance reports — dedicated views showing where your pages surface inside generative AI features on Search, such as AI Overviews and AI Mode, plus the equivalent surfaces in Discover. On 29 July 2026, Google completed the global rollout of platform properties, a new property type covering Instagram, TikTok, X and YouTube, and published a guide for analysing that data. The social rollout had been creeping out since 7 July; the experiment behind it first appeared inside Search Console Insights back in December 2025.
Read on their own, each looks like a modest feature release. Read together with everything else Google has quietly added to the tool — query groups, a branded-queries filter, custom annotations, weekly and monthly views, AI-assisted report configuration — the direction is unmistakable. Search Console is being rebuilt from a webmaster diagnostic into a visibility measurement layer that spans your site, your presence inside AI answers, and your social content. The tool changed category. Most reporting templates haven’t noticed.
Google is the one publishing the numbers now. Until this year, any claim about your visibility inside AI answers or about social content ranking in Search came from a third-party estimator. Now there is a first-party source — incomplete, but first-party. That changes which arguments you can actually win in a client meeting.
What Do the Generative AI Reports Actually Tell You — and What Don’t They?
Set expectations before you open it, because the gap between what people assume this report contains and what it contains is wide. At launch it gives you impressions only, broken down by page, country, device and date. There are no clicks, no CTR, and no query data. You can see that a page appeared inside an AI feature; you cannot see whether anyone clicked through, or what someone asked to trigger it. Google has said more metrics will follow, without committing to a date. Access also started as a subset of properties, with the UK first — so if the report isn’t in your account yet, that is expected, not a misconfiguration.
That sounds like a thin dataset, and it is. It is also the most useful thin dataset you’ve been given in years, because it answers the question everyone has been guessing at: is my content being used to build AI answers at all? A page with heavy AI impressions and flat organic clicks is being read and summarised without being visited — the zero-click pattern, now with a number attached. A page with no AI impressions is invisible to the surface that increasingly sits above the classic ten links. Those are two completely different problems, and until June you couldn’t tell them apart. If you want the strategic side of that — how to earn the citation rather than just measure it — that is the argument in Google AI Mode is the default now.
The honest caveat: an impression is not a business outcome, and nobody should be reporting AI impressions as a KPI to a CFO. Treat it as a diagnostic signal, not a result — the same way you treat impressions in the classic performance report. The mistake to avoid is building a dashboard around it before Google has finished the metric set, then having to explain a step-change in the chart when clicks eventually get added.
Search Console stopped measuring your website and started measuring your presence in Google — across pages, AI answers and social posts. Those are three different visibility problems that used to need three different tools and a lot of guessing.
The data is incomplete. It is still the only first-party data that exists.
How Do Platform Properties Work for Instagram, TikTok, X and YouTube?
This is the genuinely new one, and the part almost nobody has set up. A platform property is the first Search Console property you can verify without owning a domain — you connect a social or video account instead. Once connected, you get a Performance report showing clicks, impressions, average CTR and average position for that account’s content, spanning Google Search, Discover and Google News. Not engagement inside Instagram. Not TikTok’s own analytics. How your posts perform as search results.
Access comes two ways. Large accounts are enrolled automatically through a Search profile — the published thresholds are roughly 100,000 followers on YouTube, Instagram or X, or 300,000 on TikTok. Below that you connect the account manually. Google’s own guide walks through five ways to use the data: reading search intent through query groups, catching trending content with a 24-hour filter, exporting to compare platforms side by side, using annotations to tie changes to content edits, and comparing formats — short-form against full-length, playlists against standalone videos.
The strategic consequence is bigger than the feature. Your social content has been competing in Google search results for years, and you have been running social and SEO as separate disciplines with separate reports and, usually, separate people. This connects them with shared metrics. A Reel that ranks for a commercial query is doing SEO work, and now you can prove it, name the query, and decide deliberately whether to answer that query with a post, a page, or both. That is a planning conversation you couldn’t have with evidence a month ago.
Is your reporting still measuring 2023’s version of Google?
I help B2B teams and consultants rebuild their measurement setup around how discovery actually works now — organic, AI surfaces and social in one framework, with numbers you can defend.
What Should You Change in Your Reporting Now?
Four things, in order of effort. First, connect the platform properties this week. Historical data only accumulates from the moment you connect, so every week you wait is a week of baseline you will wish you had when someone asks whether the social push moved search visibility. This is a fifteen-minute task with a compounding cost for delay.
Second, check whether the generative AI report exists in your properties and, if it does, take a manual snapshot now. Note which pages carry AI impressions and which don’t, before the metric set changes underneath you. Third, separate the two failure modes in your analysis: pages losing clicks while gaining AI impressions need better citation-worthiness and a rethink of what the click was worth; pages with neither need the classic work. Same drop in traffic, opposite prescriptions.
Fourth, stop reporting organic and social as separate universes. If a client’s TikTok is pulling Search impressions for queries their landing pages should own, that is one finding, not two channel reports that never meet. The unified view is the actual product here — the individual reports are just the plumbing. It is the same discipline as choosing the right measurement method for the right question, which is the case made in MMM vs attribution vs incrementality.
One closing caution: none of this is a reason to rebuild your dashboards this quarter. The AI report is deliberately incomplete and will change; platform properties are new enough that best practice doesn’t exist yet. The right move now is to turn everything on, start collecting, and stay boring about the conclusions until there is enough history to say something defensible. The teams that will look prescient in six months are simply the ones who connected the accounts today.
Want a measurement setup that reflects how people actually find you?
I work with consultants, agencies and B2B teams to connect the right properties, build reporting that covers organic, AI surfaces and social together, and turn it into decisions instead of another dashboard nobody opens.
