Someone on the marketing team finally remembers to Google the company's own product name, sees an AI Overview at the top of the results with a competitor cited and the company nowhere in it, and asks why nobody flagged this sooner. The honest answer, most of the time, is that nobody was looking, because Google doesn't exactly hand you this data on a plate the way it hands you a keyword ranking.

It still doesn't, at least not directly. Here's what Search Console actually tells you about AI Overviews and AI Mode as of October 2026, what it still doesn't, and the manual and tool-based checks that fill the gap.

What counts as an AI Overview mention versus an AI Mode mention

Quick definitions, because the terms get used loosely. AI Overviews are the AI-generated summary boxes that can appear directly in regular Google Search results, above or among the normal blue links, for queries Google decides warrant one. AI Mode is a separate, fully conversational search experience, its own tab or mode in Search, built for multi-step and follow-up questions rather than a single query and a results page. Both can cite your pages. Both are, per Google's own documentation, evaluated using the same core retrieval and ranking signals as regular Search: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary," beyond the standard technical requirements for Search (Google Search Central). If your pages aren't indexed or crawlable in the first place, nothing below will matter; run through the technical SEO checklist first if that's in any doubt.

What Search Console actually shows you now

Start with the uncomfortable fact: Search Console does not give AI Overviews their own report. Google's documentation says traffic from AI Overviews and AI Mode is "included in the overall search traffic" and reported in the Performance report under the Web search type (Google Search Central). So an impression or click your page earns from an AI Overview is real data in Search Console, but it sits mixed in with your classic blue-link results.

What you can do is read the blended numbers with that in mind. Watch pages and queries where impressions rise while clicks and average position barely move; that pattern often shows up when a page starts appearing as a source in an AI Overview for a query where it doesn't rank in the top few organic spots. It's a hint, not proof, and the only way to confirm it is to look at the results page yourself.

Here's the limitation that trips people up and that matters most for anyone trying to turn this into a business case: there's no click data, no CTR, and no query-level breakdown in this report. You can see that a page got, say, 400 AI Overview impressions last week. You cannot see what query triggered them, and you cannot see whether any of those 400 impressions turned into an actual visit. Google's standard Performance report, filtered to the regular Web search type, still shows clicks and queries for ordinary search results, and sessions that start from AI Mode get folded into that same overall Web search type reporting rather than broken out separately (Google Search Central). So you end up with two incomplete pictures that don't quite overlap: an impressions-only view of AI feature visibility, and a clicks-and-queries view of everything else, with no single report connecting a specific AI Overview impression to a specific resulting click.

Treat the AI performance numbers as a visibility signal, the way you'd treat a brand-awareness metric, not as a traffic report. A page showing strong AI Overview impression growth is getting seen by a real, specific chunk of searchers. Whether that translates into visits, leads, or revenue is a question this report cannot answer on its own, and claiming it can to a stakeholder who then asks for the attached revenue number is how this feature gets a reputation for being useless that it doesn't fully deserve.

Manual checks: what to actually vary

Because Search Console can't separate AI Overview appearances from regular results, you don't know from the data alone which searches showed an AI Overview that cited you, or one that cited a competitor instead. Manual testing fills that gap the same way it does for ChatGPT, and the same variability caveat applies: AI Overviews aren't guaranteed to show for the same query every time, and the exact content of one can shift between runs.

Location changes what you see, sometimes dramatically. AI Overviews draw on location signals the way regular local results do, so a query run from a US IP will not reliably match what the same words return for a searcher in another country, and even city-to-city differences can matter for anything with local intent. If you're checking visibility for a geographically specific business, test from the actual markets you care about, not just wherever your office happens to be. A VPN or, more reliably, the country filter in Search Console's Performance report, helps here.

Log out, or use an incognito or private window, for your baseline checks. A logged-in Google account carries search history and personalization signals into the results, which skews what you see toward your own prior behavior rather than a neutral view.

Vary the query the same way described for ChatGPT testing: a direct brand query, a no-brand category query ("best [category] for [use case]"), a comparison query, and a problem-first query that doesn't name a category at all. AI Overviews show up far more often for some query shapes than others, informational and comparison-style queries trigger them more reliably than narrow navigational ones, so a prompt set skewed toward brand-name searches will systematically undercount how often you're actually missing from the Overviews that matter.

Check whether AI Mode returns a different answer than the AI Overview on the regular results page for the same query. They're built differently and don't always surface the same sources, so a page absent from one can still appear in the other.

Check on both mobile and desktop if you have the patience for it. AI Overviews have historically triggered at different rates and with different formatting across device types, and since most commercial search traffic skews mobile, a desktop-only testing habit (easy to default into, since that's where most people do SEO work) can miss the experience most of your actual buyers are having.

Preferred Sources, Highly Cited, and what Search Console won't tell you about either

Google added two labels in May 2026 that sit on top of the basic citation mechanism. Preferred Sources lets a searcher choose the sites they want to see more of, and those picks are now labeled when they appear in AI Overviews and AI Mode. The expanded Highly Cited label marks articles that many other outlets reference. Google said people had already selected more than 345,000 unique sources (Google). Neither label shows up anywhere in Search Console: there's no report and no field telling you how many searchers picked your site or which of your pages carry the Highly Cited badge.

This matters when you read a jump or drop in impressions. A change in how many searchers prefer your site, or a page gaining or losing a Highly Cited label, looks exactly like any other ranking shift in the data. Manual spot-checking is the only way to tell these apart, because the results page is the only place the labels are visible.

Check your own snippet controls first

Check your own controls before you blame the algorithm. Google documents nosnippet, data-nosnippet, max-snippet and noindex as the ways to limit what Search, including its AI features, can show from a page (Google Search Central). There's no separate Search Console switch for AI features. If a page suddenly stops appearing as a source, look for a snippet directive a developer or plugin added before you spend time on content theories.

What to log

Keep a sheet, the same discipline as the ChatGPT tracking method in our companion guide on checking ChatGPT mentions. Columns worth keeping: date, query, location/country tested from, logged-in state, whether an AI Overview appeared at all (a meaningful number of queries still return none), whether your domain was cited, the position of your citation among others listed, which competitors were cited alongside or instead of you, and whether the same query triggered AI Mode differently. Over a few months this sheet answers a more useful question than any single screenshot: not "were we ever cited" but "for which query shapes, and how consistently."

Cross-reference what you log manually against Search Console's Performance data for the same pages and queries. If your manual spot-checks show citations for a page but Search Console shows flat or declining impressions, that's worth digging into, since it can mean the citations you're catching by hand are for long-tail queries too low-volume to move the aggregate number, or it can mean your manual sample isn't representative of the real query mix.

Rank trackers that record AI Overview presence

Manual checking doesn't scale past a handful of queries without eating real time, and it says nothing about how you compare to competitors across a large keyword set. A few established SEO platforms have built AI Overview tracking into their existing rank-tracking tools, which is a meaningfully different approach from the AI-mention-monitoring tools built specifically for the ChatGPT and multi-engine category.

Semrush's position tracking and its dedicated AI Visibility Toolkit both report on AI Overview presence and citations as part of the broader rank-tracking workflow, covering Google AI Overviews and AI Mode alongside ChatGPT, Gemini, and Perplexity, with visibility, competitor, and sentiment reporting in one module (Semrush).

Ahrefs runs a comparable feature through its Brand Radar tool and a free AI Overviews tracker, built on top of Ahrefs' existing database of search queries rather than synthetic prompts; it tracks mentions, citations, and estimated impressions, weighted against real search volume for the queries involved, and the free tracker is a reasonable entry point before committing to the full platform (Ahrefs).

Both of these differ from tools like Otterly.AI and Peec AI in one important way: they're anchored to real, existing keyword data your business already ranks or competes for, so the "prompts" being tracked aren't hand-picked by you, they're drawn from actual search volume. That's a strength if you want coverage across your whole keyword universe without building a prompt list by hand, and a limitation if your buyers' actual questions don't look much like the keyword phrases these tools are built around (a conversational "what should I use for X" doesn't always map cleanly onto a keyword database built for ten-blue-links search behavior). Check what's included in your current Semrush or Ahrefs subscription before buying a separate AI-mention tool; you may already have partial coverage and not know it. Our Semrush versus Ahrefs comparison covers the broader tradeoff between the two platforms if you haven't picked one yet.

A 20-minute weekly routine

You don't need a dashboard to start. You need a fixed list and a calendar reminder.

Pick 15 to 25 queries that matter to revenue: your brand name, your brand plus "review" and "pricing", your main product category, and the comparison searches where buyers weigh you against a named competitor. Keep the list the same every week. Changing it constantly makes the trend line meaningless.

Every Monday, run each query in a private window with location set to your main market. For each query, log whether an AI Overview appeared and, if it did, which domains it cited. Your own domain either shows up in that list or it doesn't, and the rest of the list is your competitive picture. That takes about a minute per query once you're used to it.

Once a month, compare that sheet against Search Console. Filter the Performance report to the same queries and look at impressions and clicks over the same weeks. If your manual log shows you newly cited on a query and Search Console shows impressions climbing for the matching page, you have two independent signals pointing the same way. If they disagree, trust neither yet and widen your query sample.

The competitor column is the part most teams skip, and it's often the most useful. A domain that keeps appearing in AI Overviews for your category is telling you what Google's systems find citable right now. Read those pages. Usually the gap is plain: they answer the question in the first paragraph, and you answer it after three paragraphs of preamble.

Google's other AI surface is Gemini, which grounds its answers in the same search index but shows citations differently. Our guide on getting your brand mentioned in Google Gemini covers how to check and improve visibility there.

If manual checks take too long, several rank trackers now log AI Overview appearances automatically. We compared them in our guide to the best Google AI Overviews trackers.

Why track citations at all if they don't send clicks? Because most of the value now sits in the answer itself; our explainer on zero-click search covers what the click studies actually found.

This is getting more urgent: AI Overviews now appear on around 80% of branded searches, up from about 26% at the start of September.

AI Overview clicks still land in GA4 as organic search, but other AI assistants are now separable: see how to track Gemini referrals in GA4.

FAQ

Does Search Console show which search query triggered an AI Overview citation?

No. As of October 2026, Google counts AI Overview and AI Mode traffic inside the normal Performance report (Web search type) and doesn't break it out. You can see queries and pages, but not which of those impressions came from an AI Overview. That gap is exactly what manual testing and third-party rank trackers fill.

Can I filter Search Console to AI Overviews only?

Not according to Google's current documentation (Google Search Central). AI Overview and AI Mode data is blended into the Web search type. If you see a third-party guide describing a separate AI report, check it against Google's own docs before you build reporting around it.

Do I need to be logged out to get an accurate read on whether I'm cited?

For a neutral baseline, yes. A logged-in account carries search history and personalization into the results. Test logged out or in a private window for your baseline, and separately, if it matters to your use case, test logged in to see the personalized experience your repeat visitors might get.

Can AI Overviews cite a page that doesn't rank in the top results for that query?

Yes. Citation selection in AI Overviews runs a separate evaluation pass on top of normal retrieval, and independent analysis has repeatedly found pages ranking outside the top five organic results still getting cited, because the synthesis stage rewards how cleanly a passage can be extracted and attributed, not just how the page ranks. See our full breakdown of how AI Overviews pick citations for the mechanics.

Is Ahrefs or Semrush better for tracking AI Overview presence specifically?

Both track it, built on top of their existing rank-tracking and keyword databases rather than as a standalone product. Which one serves you better usually comes down to which platform your team already uses for regular rank tracking, since the AI Overview data sits inside that existing workflow rather than as a separate purchase decision for most plans. Check current feature inclusion on your tier before assuming you have it.

How do I know if I've been designated a Preferred Source or given the Highly Cited label?

There's no report for either one. Both are visible only inside the actual AI Overview or AI Mode answer where they apply, which means the only way to catch them is the manual query testing described above, run against the specific topics where you'd expect that kind of consistent treatment. A sustained pattern of citation across many related queries in one topic area is the closest indirect signal you'll get without seeing the label itself.

Pull up Search Console's Performance report this week and sort your pages by impressions gained over the last three months. Pages whose impressions climbed while clicks stayed flat are your first candidates: search their top queries by hand and see whether an AI Overview is citing them. Everything else in this piece is about finding the queries where you should be in that answer layer and aren't yet.