SERP Checkers and AI Overviews: What Gets Captured, and What Doesn't

A SERP checker sends a query to Google, receives the results page, and parses it. When an AI Overview is present, the checker pulls it out as its own item.

As of October 2026, a typical SERP API returns three things for the overview. DataForSEO, for example, documents an ai_overview item with these fields:

• The overview text, often as markdown, so you can read exactly what Google said.
• Structured items, the blocks the overview is built from.
• References, each with a URL, domain, title and a short text excerpt. These are the supporting links shown beside or under the overview.

That is the whole capture. It tells you what Google displayed and which pages it linked. It does not tell you which pages Google retrieved and dropped, what queries it ran behind the scenes, or why one source beat another.

One technical detail matters more than it looks. Some overviews load asynchronously, after the first page response. A checker has to make a follow-up request to catch them. If your tool skips that step, it can log "no overview" for a query where users see one. Ask your tool vendor whether it captures asynchronous overviews.

There's no official Google API for AI Overviews, which is why tools rely on SERP capture. Our explainer on how tools track AI Overviews without an API covers the methods in more depth.

Google says plainly that AI Overviews often do not trigger. An empty result is not a broken checker. It is the most common outcome for many query types.

This changes how you should read a report. Say you run a dental clinic in Austin and track 50 keywords. If 12 return an overview, your visibility question only applies to those 12. The other 38 are classic SERPs, and the old rules for organic results and featured snippets versus AI Overviews still decide what users see.

What to do with this:

• Split your keyword list into "overview present" and "overview absent" before you measure anything else.
• Report citation counts against queries that had an overview, not against your full list. Otherwise a quiet week for overviews looks like a drop in your performance.
• Re-check on a schedule. A single capture cannot tell you whether a query triggers an overview reliably or only sometimes.

If you want a quick first pass, the free check for which keywords trigger AI Overviews sorts a list this way.

Location and language are inputs to the request, not settings for display. A SERP checker asks Google for a page as if it came from a specific place in a specific language. Change either, and you can get a different SERP, a different overview, or no overview at all.

This creates three practical rules:

1. Match the market you sell to. A US-English capture says little about your UK or German buyers. Set location and language per market, and track them as separate keyword sets.
2. Go local when the business is local. For a clinic, a restaurant or a law firm, a city-level location is closer to what buyers see than a country-level one.
3. Record the settings with every result. A citation captured in one market and missing in another is not a contradiction. It is two different SERPs.

AI Mode is a separate surface again. Google states that AI Overviews and AI Mode may use different models and techniques and return different links. SERP providers expose AI Mode through a separate endpoint. Check it separately, for example with the free AI Mode checker, and never merge the two into one number.

A SERP capture is a measurement of what Google showed. It is not a window into how Google decided.

No position to track

There is no rank or position inside an AI Overview that anyone can verify. Google's Search Console report for AI features shows impressions only, with no clicks, CTR or position, and Google says third-party tools don't have access to its internal ranking data. Treat any "AI Overview rank" as the tool's own construct.

No view of the real queries

Google may use a query fan-out technique, issuing several related searches across subtopics. Your keyword is the user's input, not necessarily the query that retrieved your page.

Cited is not the same as used

A reference link shows Google displayed your page. It does not prove the overview's wording came from it, or that an uncited page had no influence.

Organic rank is a weak proxy

Ahrefs found that 37.9% of URLs cited in AI Overviews also appeared in the first 10 results (March 2026). That leaves most citations coming from outside the top 10, so a strong rank report does not stand in for overview capture. Ahrefs changed its method between studies, so don't read older figures as a trend.

Google's eligibility rule is short. To appear as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Search with a snippet. Google says there are no additional technical requirements and no special optimizations necessary. Eligibility is not inclusion, though: Google also says indexing and serving isn't guaranteed.

Two access traps are easy to miss. First, snippet controls cut AI display too. A max-snippet:0 or data-nosnippet copied from a template can silently keep a page out of overviews. Second, Search Console has a generative AI inclusion control. Inclusion is the default, but child properties inherit the parent setting, so check it once.

On content, the only positive signal in the evidence is simple: a specific fact (a price, a number, a one-line definition) stated on a page Google can fetch gives the overview a sentence it can quote. For how this plays out on Google specifically, see getting cited in AI Overviews.

Run these in order. Each step makes the next one meaningful.

1. Build the keyword list by market. Choose the questions buyers actually ask, then set location and language for each market you sell in.
2. Sort by trigger. Capture every query and mark which ones returned an overview, including asynchronous ones. Only overview queries move to step 3.
3. Record references, not just presence. For each overview, store the cited URLs and domains. Note whether your pages appear, and which other sites do.
4. Read the text for accuracy. The overview can name you and still get your price, location or offer wrong. Compare the captured text to your own facts, as described in checking whether an AI Overview describes you accurately.
5. Check access for missing pages. Where you expected a citation and got none, confirm the page is indexed, snippet-eligible and not excluded by a stray directive or the Search Console inclusion setting.
6. Note third-party sources. When overviews cite review sites, directories or publications instead of you, being named accurately on those pages is the route into the answer. That's an association in the evidence, not a guaranteed effect.
7. Repeat on a schedule. Compare like with like: same query, location and language.

For a one-off look at a single query, the AI Overview Checker on 99Visibility shows what Google displays and which sources it cites.

• Treating "no overview" as a loss. It is often just a query that doesn't trigger one. Measure citations only where an overview exists.
• Checking from the wrong place. A default US-English capture misleads any business selling elsewhere.
• Blending surfaces. AI Overviews, AI Mode and the Gemini app are different products. A single blended score hides which one is failing.
• Reporting a rank. No ground truth exists for position inside an overview. Report presence, references and accuracy.
• Adding markup to get cited. Google says structured data isn't required for generative AI search. Ahrefs' matched test found AI Overview citations fell 4.6% after pages added schema. Existing markup does no harm, but it isn't a lever here.
• Trusting one capture. A snapshot shows one request. Decisions need repeated, consistent checks.

For a wider view of what different checkers can and cannot measure, read AI SEO checkers: what a check can tell you.

Why does my checker say there's no AI Overview when I saw one? Three common causes. The overview may load asynchronously, and the tool didn't make the follow-up request to capture it. Your location or language may differ from the checker's settings. Or the query triggers an overview only some of the time, so one capture missed it. Confirm the tool's location, language and asynchronous handling, then re-check before drawing conclusions.

Does Google Search Console show AI Overview data? Partly. As of October 2026, Search Console's report for AI features shows impressions for AI Overviews and AI Mode, with dimensions like pages, countries, dates and devices. It shows no clicks, CTR or position, and Google offers no API for it. It tells you that your pages appeared, not how often they were clicked or what the overview said about you.

Can a SERP checker tell me my ranking inside an AI Overview? No tool can verify one. Google publishes no position data for AI Overviews, and it states third-party tools don't have access to its internal ranking data. A checker can record whether your page is among the cited references and in what order they were displayed in that one capture. Treat display order as an observation, not a rank.

Is checking AI Mode the same as checking AI Overviews? No. Google states AI Overviews and AI Mode may use different models and techniques and can return different links. SERP providers expose them through separate endpoints. A page cited in an overview may be absent from AI Mode for the same query, and the reverse. Track them as separate surfaces and report them separately.

Does adding schema markup help my page appear in AI Overviews? The evidence says it isn't a lever. Google says structured data isn't required for generative AI search and that no special optimizations are needed. Ahrefs' matched-control study found AI Overview citations fell 4.6% on pages that added schema. Leaving existing markup in place does no harm. Focus instead on indexing, snippet eligibility and clearly stated facts.