Search Console to show AI impressions for UK sites
Native AI impression data from Google ends two years of guesswork and shifts the budget argument for any brand producing citation-ready content.
Key takeaways
- Google is testing AI Search impressions inside Search Console for select UK sites.
- Brands will be able to count AI citations against the queries that triggered them, for the first time natively.
- Authoritative explainers from banks, multilaterals and industrial groups are likely to look stronger in AI data than in traffic data.
- Third-party AI citation trackers face a free, more accurate competitor from Google itself.
- Waiting for the rollout before investing in citation-ready content cedes the head start.
Google has begun showing a small group of UK publishers something marketers elsewhere have spent two years guessing at: how often their pages surface inside AI Search answers. Search Engine Journal reports that Google is testing dedicated Search Console controls and AI-specific impressions data, with select UK properties already seeing the new view.
The detail matters more than the launch. Until now, the gap between "ranking on Google" and "being cited by Google's AI" has been a black box. SEO teams have inferred AI visibility from third-party crawlers, Bing's slightly more generous data, and the occasional screenshot. Google's own telemetry, if it ships broadly, ends the guesswork for the one surface that still drives the majority of B2B research traffic.
What the test actually exposes
Two things, according to SEJ's reporting. First, controls: site owners appear to get levers over how their content is used in AI Search, which implies an opt-signal Google can point regulators toward when the inevitable copyright question arrives. Second, reports: AI-specific impressions, separated from the standard blue-link tally.
That second piece is the load-bearing one. An impression inside an AI Overview or AI Mode answer is not the same unit as an impression on a results page. It means a model retrieved your page, judged it relevant, and surfaced it (as a citation, a quoted line, or grounding context the user never sees attributed). For the first time, brands will be able to count those events against the queries that produced them.
Who this rewards
Publishers with strong topical authority on narrow questions will look better than their traffic numbers currently suggest. Brands that have been quietly cited inside AI answers without earning a click, the most common complaint of the past 18 months, finally get a number to put next to the lost traffic. Expect a wave of internal decks in Q1 reframing "organic decline" as "AI impression growth," some of it justified.
For financial services and multilaterals, the implication is sharper than for consumer brands. A central bank explainer, a UN agency's methodology note, or a ratings firm's sector primer tends to be exactly the sort of authoritative, structured content LLMs prefer to ground on. These organisations have long suspected they are over-indexed in AI answers relative to their search traffic. Search Console data will either confirm it or puncture it. Either result changes the budget conversation.
Industrial groups face a different read. Product pages, spec sheets, and sustainability disclosures are cited unevenly by AI systems, and the variance is rarely visible. A Holcim or a Siemens will be able to see, query by query, whether its own technical content or a competitor's is the source the model leans on. That is the first piece of competitive intelligence on AI visibility that does not require buying a third-party tool.
The caveats worth pricing in
This is a test, on a subset of UK sites, with no announced rollout date. Google's track record on Search Console features is to ship slowly and rename twice. The impressions metric will also need definition: does a citation in AI Mode count the same as a passage quoted without a link in an AI Overview? If Google conflates them, the number will flatter more than it informs.
There is also a strategic question Google has not answered. Giving sites controls over AI Search usage invites the same publisher revolt that hit OpenAI and Perplexity, except Google cannot afford mass opt-outs without degrading its own answers. The controls, when they arrive, will likely be narrower than publishers want.
What changes for brands now
Three things, none of them theoretical. Communications teams that have been writing AI visibility into 2026 plans without a measurement layer now have one coming. The internal argument for producing structured, citation-ready content (clear claims, named sources, dated figures) gets easier to win when the CFO can see the impression count. And the agencies and tools that have built a business on inferring AI citations from sampling now face a free, native, more accurate alternative.
The brands that benefit first will be the ones already producing the kind of content LLMs prefer to cite. The ones who waited for proof before investing will discover that proof and head start are not the same thing.