How Will Google's AI Click Reporting Affect Your Brand Strategy?

What changed in Google's AI click reporting?

Google has not yet released a dedicated Search Console report that separates clicks or click-through rate from AI Overviews and AI Mode. The current Generative AI Performance report combines impressions from those has and lets site owners review them by page, country, device, and date. It does not list clicks, click-through rate, or queries.

That leaves a measurement gap. Google counts a click on a link in AI Mode or an AI Overview as a click, but that activity is included in the wider web Performance report. A brand can see total web clicks, and it can see combined AI-has impressions, yet it cannot use Search Console's current dimensions to isolate the clicks produced by those AI features.

The distinction matters because an AI answer can surface a brand without sending a visit. A source link may earn attention, establish preference, or put a company into a shortlist before a prospect reaches its site. Traffic remains important, but it cannot be the only signal used to judge visibility in answer engines.

Search Engine Journal reported that Google said it had nothing to announce about adding more AI Search Console data. The report followed discussion of possible click and query reporting, but Google's Gary Illyes did not confirm a release or timetable. Treat separate AI click reporting as an unresolved product and regulatory question, not as a has that is already available.

The practical change is therefore not a new dashboard. It is a sharper limit on what the existing dashboard can prove. Search Console can support visibility analysis, but it cannot presently assign a clean web click total to AI Overviews or AI Mode.

1. Before: Search Console can show combined AI Overview and AI Mode impressions in its Generative AI Performance report.
2. Now: clicks from those has count in the standard web Performance report, without a dedicated AI search appearance dimension.
3. What to do: keep Search Console data, then add a prompt-level citation measurement process so visibility does not disappear inside aggregate web totals.

Who does the reporting gap affect?

The reporting gap affects any brand that expects prospects to ask Google for recommendations, comparisons, local options, or instructions. B2B software teams face it when a buyer asks for the best tool in a category. Multi-location businesses face it when a customer asks for a service in a specific city. Publishers face it when Google uses their work as source material in an AI answer.

It also affects teams that report organic search performance as though every impression and click comes from a conventional result page. That assumption was already weak when search results included rich results and knowledge panels. AI search makes it weaker because the answer itself can meet part of the user's need before any site visit occurs.

Marketing leaders should avoid treating the gap as proof that AI search has no commercial effect. A missing referral breakout means the effect cannot be cleanly measured through the current Search Console interface. It does not mean a cited brand was absent from the decision process.

The same caution applies in reverse. An AI impression is not proof of a visit, lead, or revenue event. Google defines an impression as a user seeing, or potentially seeing, a link to a site under its reporting rules. An impression should therefore be read as evidence of availability in the answer environment, then paired with page, conversion, and sales data where those connections can be responsibly made.

For publishers serving UK users, the Competition and Markets Authority's publisher conduct requirement raises the stakes. Its rules call for clear and detailed user-engagement metrics for publisher content used in Google AI search features, including impressions, clicks, and click-through rate aggregated across AI Overviews and AI Mode. The requirement concerns Google's conduct in the UK. It does not automatically create a global Search Console has for every brand.

That geographic boundary still matters to global teams. UK implementation can reveal the reporting model Google considers workable. It can also move a conversation that has largely relied on estimates toward a defined set of publisher metrics. Do not promise that outcome until Google publishes the actual reporting change.

1. B2B SaaS teams measuring category and competitor prompts.
2. Local businesses measuring visibility for service and location questions.
3. Publishers whose work appears as source material in AI search.
4. SEO teams whose reports combine AI-has exposure with conventional web-search totals.

Why are impressions alone an incomplete brand metric?

Impressions alone are incomplete because they show that a link appeared, not whether an answer cited the brand prominently, whether the user selected it, or whether the brand won consideration. They are still useful. They establish that Google exposed a page in an AI feature. The mistake is treating that signal as a complete funnel.

A brand strategy needs a second measurement layer that starts with the answer. Track the prompts that is real buying, comparison, service, and research questions. Record whether the brand is cited, which page is cited, which competitors appear, and whether the answer names a competitor instead. This is the basis for Citation Share: the percentage of relevant AI answers in a category that cite a brand.

Citation Share makes the reporting gap visible rather than invisible. Search Console may show an increase in AI impressions while a tracked prompt set shows that competitors receive more citations in category recommendations. The reverse can also happen. A brand can gain citations for high-intent answers before aggregate reporting makes the shift obvious.

Keep the measures separate. Citation Count per day measures volume. Answer Presence measures breadth across a question universe. Share of Voice measures a brand against named competitors. Website sessions and conversions measure on-site behavior. Combining them into one unqualified number makes it harder to explain what moved and why.

This is also where content quality enters the strategy. Google and other answer engines need clear, current sources to cite. A page designed for a real question, with a direct answer, dated evidence, explicit scope, and helpful supporting detail gives the engine a more usable source than generic campaign copy. That is Citation Engineering: building coverage and authority that makes a brand more citeable without claiming to control the model.

Do not respond by chasing only pages that have already appeared in Search Console. That can narrow coverage to yesterday's visible questions. Use current visibility data to find proof of demand, then map adjacent questions, comparisons, and local variations where the brand needs a credible source before a competitor owns the answer.

1. Measure AI-has impressions in Search Console.
2. Track Citation Share across a defined prompt set.
3. Review cited pages and competing domains for each material prompt.
4. Compare AI visibility with qualified traffic and conversion evidence, without assigning causation that the data cannot support.

How should your brand respond before separate AI click data arrives?

Your brand should build an answer-level baseline before Google changes its reporting. Start with the questions that lead to a real commercial decision. Use category, comparison, implementation, location, and problem questions that a buyer or customer would plausibly ask. Keep the list stable enough to compare over time, then add new questions through a documented process.

For every prompt, capture the engine, date, answer text needed to identify the result, cited domains, cited URLs, and whether your brand was cited. Record a citation only when the answer actually cites or links to the brand's content. Do not turn an unlinked mention into a citation. This keeps Citation Share comparable across reporting periods.

Next, inspect the cited page. Ask whether it answers the question directly, identifies who the information applies to, gives dated evidence where a number is used, and supports the claim with source links. If a page cannot meet those checks, publishing more variations of the same weak page will not solve the problem.

Then connect the audit to Search Console. Review the Generative AI Performance report by page and date. A growing AI-impression pattern can help prioritize the pages that are already being exposed. The standard Performance report can still show web clicks and conversions from search, but label those figures correctly: they include activity beyond AI Overviews and AI Mode.

Use first-party analytics carefully. Landing-page trends, assisted-conversion evidence, sales-call notes, and self-reported discovery can add commercial context. They cannot identify an AI click unless the referral and event data actually does so. The reporting standard should state the source, definition, and limitation beside every conclusion.

Finally, prepare for a future breakout without redesigning the whole reporting system. Create fields for AI Overview clicks, AI Mode clicks, AI click-through rate, and publisher or page-level attribution. Leave them empty until a documented source provides the data. A prepared schema lets a team adopt new reporting without backfilling estimates as facts.

1. Define a prompt universe tied to real buying and service questions.
2. Measure citations and competitors at the answer level on a repeatable cadence.
3. Improve the pages that need to become citable sources.
4. Use Search Console AI impressions as a visibility input, not an AI-click total.
5. Keep future AI click fields empty until Google provides verified data.

What should leaders ask for in the next reporting cycle?

Leaders should ask for a report that separates what is observed from what is inferred. The observed layer should include Search Console AI-has impressions, answer-level citations, cited URLs, competitor citations, standard web clicks, and first-party conversion events. The inferred layer should be limited to clearly labelled hypotheses about how visibility may contribute to demand.

Ask whether the brand's prompt coverage reflects its market. A software firm that tracks only its own name will miss the category questions where a buyer creates the shortlist. A service business that tracks one city will miss the location modifiers that shape local demand. Coverage should reflect the questions the market asks, not only the terms the brand prefers.

Ask which pages earn citations and which pages lose them. This turns reporting into an editorial backlog. A page that appears but does not answer a core question may need a tighter opening and better evidence. A competitor page cited repeatedly may reveal a comparison, definition, or practical guide your site has not supplied.

Ask whether the report distinguishes citation presence from traffic. A cited source can help a brand get chosen even where the user does not click immediately. A click can also originate from a standard result. Both are meaningful, but neither should be substituted for the other.

The current gap is inconvenient, but it gives brand teams a useful discipline: measure what the engines display, measure what users visit, and keep the two datasets honest about their boundaries. When Google publishes a separate AI click breakout, it can improve this system. It should not be the first time a brand looks at whether it gets cited.

1. Which decision prompts does the brand need to win?
2. Where is the brand cited, and where do competitors receive the citation?
3. Which source pages answer those prompts with direct, current evidence?
4. Which figures are observed in a platform, and which conclusions remain hypotheses?

Sources

Source: Competition and Markets Authority

The UK CMA publisher conduct requirement sets expectations for detailed engagement metrics for content used in Google AI search features. Competition and Markets Authority, 2026-08-03

Source: Search Engine Journal

Google said it had no announcement about additional AI Search Console data ahead of the reported UK compliance date. Search Engine Journal, 2026-11-03