Tracking AI Overview Impact in Search Console

Search Console doesn't have a filter labeled "AI Overview traffic." That's the first thing to get straight before trying to track it. What follows is what Search Console actually shows, and the workarounds people use to infer AI Overview impact from data that wasn't built to isolate it.

Why there's no clean filter

Google Search Console reports impressions and clicks by query and page, split by search type (web, image, video) and, in the performance report, by search appearance in some cases. AI Overviews aren't broken out as a distinct search appearance filter the way, say, "rich results" or "AMP" once were. Google has talked publicly about AI Overviews affecting search behavior broadly, but the company hasn't shipped a first-party Search Console segment that isolates "impressions where an AI Overview appeared above your listing." Anyone claiming Search Console has this exact filter is describing a feature that doesn't exist yet.

What you can actually observe

The workaround most people use is indirect: watch impressions and click-through rate for queries known to trigger AI Overviews frequently — broad informational queries, "best," "how to," comparison-style searches — and compare CTR on those against comparable queries less likely to trigger an overview. A pattern some sites report anecdotally is rising impressions with falling CTR on exactly the query types AI Overviews tend to answer directly, which would be consistent with more searchers getting an answer above the organic results without clicking through. That's a plausible read of the pattern, not a confirmed causal measurement, since Search Console has no field that says "an AI Overview appeared here" to actually prove it.

For local queries specifically, AI Overviews and local SEO and the deeper pillar on AI Overviews for local businesses in India cover what these overviews tend to draw from — local pack data, GBP fields, and web content — which is the mechanism behind why local queries are affected at all.

Building a manual tracking sheet

Since Search Console won't isolate this automatically, a manual log is the practical fallback: pick 15 to 20 queries relevant to the business, search them directly (logged out, to avoid personalization skew) on a set cadence, note whether an AI Overview appears and whether the business is cited inside it, then cross-reference the same queries' impression and CTR trend in Search Console over that period. This won't produce a clean statistical proof of causation, but it builds a directional picture over a few months that a single Search Console pull never will.

This is the same manual-tracking limitation that applies to LLM citation monitoring generally right now — measuring AI search visibility covers the broader tooling landscape, AI citation monitoring covers the query-log approach in more detail, and it's worth being honest that this space doesn't have mature, standardized measurement yet across any of the major AI surfaces, Google's own AI Overviews included.

Segmenting by query type inside Search Console

Filter the performance report by query, sort by impressions, then look specifically at question-phrased and comparison-phrased queries versus branded or transactional ones. Question queries ("what is," "how do I," "why does") are the category Google has said AI Overviews target most, so isolating that segment and watching its CTR trend over time relative to the rest of the site's queries is the closest proxy available inside the tool as it exists today.

Filtering by page also helps — pages built around a single clear answerable question (an FAQ page, a glossary entry, a how-to guide) are more likely to be the kind of content Google's AI Overview systems pull from, so watching those pages' performance trend specifically, rather than site-wide averages, gives a cleaner signal. The local SEO KPIs most teams already track can absorb this segmentation as one more cut of the same data rather than a separate report.

Schema's role in this, and its limits

Structured data doesn't guarantee AI Overview inclusion, and no one, including Google, has published a confirmed mechanism where adding schema causes AI Overview citation. What structured data does is make content easier to parse and extract cleanly, which is a reasonable precondition for extraction even without a proven causal link to inclusion. Schema markup for AI search, structured data for AI answers, and local business schema cover the implementation side of this without overstating what it buys you.

Where GBP data feeds into this for local queries

For queries with local intent, the AI Overview mechanism often draws on the same underlying local pack and GBP data that traditional local search results rely on. Keeping that source data current, through GBP write-back that publishes edits, posts, and review replies live rather than sitting in a draft state, is the practical lever available here, since there's no way to directly petition Google's AI Overview system the way you might reach out for a listicle mention. Entity SEO and the knowledge graph and entity authority and SEO cover the broader entity-signal side of this.

What not to promise a client

Don't tell a client you can show them exact AI Overview impression counts from Search Console, because the tool doesn't isolate that number. Don't tell them a specific schema change will guarantee inclusion, because no one has that guarantee to give. The honest version is a directional read built from query-type segmentation and manual spot-checks, tracked over months rather than days, the same discipline covered in GBP performance insights for the broader reporting workflow this slots into.

Can I see exactly how many AI Overview impressions my site got in Search Console? No. Search Console doesn't currently break out AI Overview appearances as a distinct filter, so this number isn't directly available through the tool.

What's the best proxy for AI Overview impact inside Search Console? Watching impression and CTR trends for question-phrased and comparison-phrased queries over time, and comparing them against the site's other query types, is the closest available proxy right now.

Does adding FAQ schema guarantee my page gets pulled into an AI Overview? No. Schema makes content easier to extract cleanly but doesn't guarantee inclusion, and there's no published mechanism confirming a direct causal link.

Is Angryturtle's AI citation tracking live yet? Angryturtle's Share of AI Voice tracker exists as infrastructure but is not yet running live citation checks — that capability is still coming online. Manual query-based checks remain the reliable method in the meantime.

Related: AI citations across ChatGPT, Perplexity, and Gemini and AI mode in Google local search cover the surfaces adjacent to AI Overviews that face the same measurement gap.

Angryturtle keeps the GBP and entity data behind these AI surfaces current, since that's the lever you actually control →