How to track visibility in Google AI Overviews: five monitoring methods

Your page ranks well, impressions are rising, and clicks are falling. That is usually not a title or meta description problem. It is Google AI Overviews - AI-generated answers at the very top of the results - taking the click before the user reaches the organic list. The difficulty is that Google AI Overviews leave no trace in standard reports. They have to be tracked separately.

Ahrefs reports AI Overviews appearing on roughly 20.5% of queries globally, and the share is higher in some markets. The effect concentrates on informational queries - precisely the ones that build traffic to blogs and knowledge bases. This guide covers five methods, from entirely free to expert-led, for measuring the real impact of AI Overviews on your visibility.

Why AI Overviews need separate tracking

AI Overviews are a new results layer that changes the economics of organic traffic. For a share of queries the user receives a complete answer at the top of the page and never clicks a result at all. That is the zero-click effect: the position stays, the click disappears.

The key observation is that AIO does not hit evenly. Queries of the form "what is", "how does it work" and "can I" are absorbed most often, because these are exactly the questions a model can answer in a few sentences. Transactional and comparison queries lose less, for now. So the first thing worth doing is establishing how much of your traffic rests on informational phrases - those are the most exposed.

What Google Search Console will NOT tell you about AI Overviews

Search Console has no AI Overviews report. There is no "show queries with AIO" filter and no "share of impressions absorbed by AI" metric. This is not an oversight - Google has no incentive to show how much traffic its own answers absorb. That is unlikely to change quickly, so the picture has to be assembled from several sources.

The good news is that Search Console shows symptoms from which AIO presence can be inferred. There are three characteristic patterns.

Three Search Console patterns that reveal AIO

1. 1Impressions rise, CTR falls. A query collects more and more impressions while CTR drops despite a stable position - something has appeared above the organic results and is taking the clicks.
2. 2Position stable, clicks down. You have held positions three to five for months, yet clicks fall quarter on quarter - with no change in seasonality and no new competitor in the top three.
3. 3Informational queries dominate the losses. The steepest CTR declines affect "what is", "how does it work" and "can I" queries - the ones AIO answers most readily.

Method 1: manual SERP verification (free)

The simplest method: open Google in a private window, enter the query and look at whether an AI Overview block appears above the results, and which sources it cites.

1. 1Collect a list of 10 to 20 priority queries - ideally the ones showing the CTR anomaly in Search Console.
2. 2Open a private window and set the language and region you are targeting.
3. 3Enter the query and check whether an AI Overview is generated - expand it and note the domains cited.
4. 4Record the result in a sheet: query, whether AIO appears, whether you are cited, and who is cited instead of you.
5. 5Repeat on a cycle, for example weekly, to catch changes.

When this is enough: a small site, a couple of dozen queries, no budget. When it is not: hundreds of queries, a need for historical data, or a client-facing report. One caveat - AIO renders differently depending on location and device, so a private window does not guarantee a "clean" result.

Method 2: Search Console plus Looker Studio (free)

The best free starting point. Instead of reviewing Search Console by hand, build a Looker Studio dashboard that automatically flags queries with a suspicious profile: high position, plenty of impressions, low CTR.

1. 1Connect Search Console to Looker Studio as a data source using the native Google connector.
2. 2Build a table with columns for query, average position, impressions, clicks and CTR.
3. 3Add a filter: position in the top ten, at least 50 impressions, and CTR below 1%.
4. 4Add a period comparison: current CTR against CTR six months ago for the same queries.
5. 5Colour-code queries where CTR has more than halved while the position held - that is your list of AIO suspects.

The result is a ready list of queries where an AI Overview may be showing. It is a proxy only. The method cannot distinguish a decline caused by AIO from one caused by a featured snippet, so confirm the most important queries manually.

Method 3: dedicated AI Overviews monitoring tools

Once you are serving clients or managing a large site, manual checks and a dashboard stop being enough. You need a tool that detects AIO presence in results by itself and tracks it over time. There is now a reasonable choice of them.

Semrush AI Toolkit has the widest coverage - Google AIO plus ChatGPT, Perplexity and Gemini - so it suits teams thinking about GEO more broadly than Google alone. Ahrefs Brand Radar lets you check for free whether your brand appears in AI answers at all, which makes a good first step. A fuller comparison is in the guide to AI visibility monitoring tools.

Method 4: track more than Google - ChatGPT, Gemini and Perplexity

This is where conventional SEO and genuine AI visibility measurement part ways. Google AI Overviews are only one of the surfaces your buyers use. If you measure AIO alone, you have a fragment of the picture. The full picture is visibility across four environments - Google AIO, ChatGPT, Gemini and Perplexity - because together they determine whether AI recommends you or a competitor.

Alongside position, it is worth tracking Share of Model - your brand's share of AI answers to buying questions in your category. The methodology for measuring across four models is set out in the guide to measuring brand visibility in AI, and the practice for a single model in the instructions for Perplexity.

Method 5: continuous expert monitoring

If AI visibility genuinely affects your business, a one-off screenshot is not enough - you need repeatable measurement over time, compared against competitors and properly interpreted. That is what an AI visibility audit is: a manual analysis of the brand across ChatGPT, Gemini, Perplexity and Google on a set of buying questions, answering who AI recommends instead of you, and why.

The metric you are probably not tracking yet: AI Citation Rate

Most practitioners treat AIO purely as a threat - "AIO is eating my traffic". Fair enough, but that is half the picture. Your domain can also be cited as a source inside an AI Overview, and that is a measurable opportunity. AI Citation Rate is the new KPI: how often your page appears as a link inside the AI answer block itself, not only as an organic result beneath it.

What drives citations? An Ahrefs study across 75,000 brands found correlations that upend conventional thinking about links:

The implication: brand mentions and a presence on YouTube correlate with AIO visibility roughly three times more strongly than backlinks. That is because YouTube is one of the most frequently cited domains in AI Overviews, and video transcripts become source material for models. Building brand recognition beyond Google itself has stopped being "soft" marketing - it is now a direct AI visibility signal.

What to do when AI Overviews leave you out

• Answer directly: put concise, two- or three-sentence answers to specific questions at the start of a section - that is the format AIO cites most readily.
• Structure the content: question-form headings, lists, tables and structured data (FAQ and Article schema) make it easier for a model to lift a passage.
• Build brand mentions: industry publications, podcasts, social media and a presence on YouTube all increase the chance of a citation.
• Support claims with data: AI cites statements backed by figures, research and first-party case studies more readily.
• Measure citations, not only losses: add AI Citation Rate to the report alongside conventional positions.

There is more detail in the guide to optimising for Google AI Overviews and AI Mode.

Which method to choose? It depends on scale

In summary

AI Overviews are not going away - their share of queries has grown from a few per cent in 2024 to around a fifth of all queries globally, and higher in some markets. Anyone not measuring this is working blind. Start with the free methods (private-window checks plus Looker Studio), confirm manually, and reach for a dedicated tool as you scale. And remember: Google AIO is only one surface - you measure real AI visibility only when you also look at ChatGPT, Gemini and Perplexity.