How to Check If Your Business is Appearing in Google AI Overviews

Everybody wants a clean dashboard for AI Overview visibility. Preferably with tidy charts, attribution lines, and a reassuring green arrow pointing upward. Unfortunately, Google has not built that yet, which means a surprising amount of GEO observation currently involves doing something the industry tends to dislike: manually checking things yourself.

The awkward reality is that AI Overviews are generated responses, not fixed rankings sitting in a predictable position every time a query is searched. They vary by query type, location, session context, and what Google decides is worth generating in the moment. Which makes tracking them messier than traditional SEO reporting, and considerably less satisfying for people who enjoy spreadsheets behaving properly.

Google AI Overviews do not appear on all search queries. They appear most consistently on informational and research queries (“what is”, “how does”, “best [option] for [situation]”) and on comparison queries. They rarely appear on commercial or transactional queries: product searches, queries with “buy”, “price”, or “near me” in them. A query like “accountant Sydney CBD” may or may not trigger an AIO. “What should I look for in an accountant for a small business?” is much more likely to.

Before running any check, make sure the queries you are testing are the kind that actually trigger AIOs. What Does It Actually Take to Appear in AI-generated Answers? covers query types and AIO eligibility in more detail.

• Write down the three to five queries most relevant to your business: the questions your customers ask before they decide to engage, not the transactional queries they use when ready to buy.
• For each query, ask yourself: is this informational or research-oriented? If yes, it is worth testing. If it is a commercial query (“book accountant Sydney”), set it aside for now.

The goal of a clean session is to see what Google shows a user with no history: no signed-in account, no prior searches on the topic, no personalisation. This is the closest you can get to an unbiased observation.

Open a private or incognito browser window. Do not sign in to any Google account. Do not have any prior searches on related topics in this session.

A note on incognito: it is not perfectly clean. Google can still infer some context from the session, your general location for example. But for a practical check, incognito in a fresh window is clean enough to be useful. It removes personalisation and search history, which are the primary sources of variation. What it cannot remove is location.

If your business serves a specific location, test from that location. If you are physically somewhere else, this is worth noting as a limitation. A search run in Sydney will return different AIO results than the same search run in Melbourne or overseas. For local businesses, geography significantly affects what appears.

• Open a new incognito or private browsing window.
• Do not sign in to any Google account.
• Go to google.com.au (or the relevant country domain for your market).
• If you are testing for a specific location your business serves and you are not currently there, note this as a limitation of the check.

Type your first query into Google. Before looking for your business, look at what appears.

An AI Overview appears as a panel above the organic search results, usually with a coloured highlight or “AI Overview” label. It contains a generated summary paragraph or paragraphs and typically cites three to eight source pages, shown as small cards beneath the text.

If no AI Overview appears, meaning you see only organic results, ads, and standard SERP features, the query may not be AIO-eligible for this search. This does not mean you have failed a GEO check. It means this query type is not triggering an AIO, and you should adjust your query list.

Try a few variations of the query with slightly different phrasing before concluding it is not AIO-eligible. Sometimes adding “in [location]” or rephrasing from a direct question to a research phrasing (“best accounting advice for small business” versus “accountant booking”) changes whether an AIO appears.

• Run the query in your clean session.
• Note whether an AIO appears or not.
• If no AIO: try two or three phrasing variations. If still no AIO, mark this query as non-eligible and move to the next one.
• If an AIO appears: proceed to Step 3.

If an AIO appears for your query, look for your business in two places.

In the AIO body text: Does your business name appear in the generated paragraph or paragraphs? Being named in the answer itself is the strongest form of AIO citation. The AI system has identified your business as a relevant answer to the question.

In the source cards: Below the AIO body text, Google shows the pages it drew from. Your business may appear as a source card without being named in the answer body. This is also a citation, but a different one. Google has used your content as a source without surfacing your brand name prominently in the answer.

Both are worth noting. They are not the same thing. Appearing in source cards means your content is being indexed and accessed by the AIO system. Being named in the answer means your brand has been identified as specifically relevant to the query. The latter is the goal; the former is progress toward it.

• Read the AIO body text carefully. Is your business name mentioned?
• Check the source cards. Does your website appear?
• If you appear: note which query triggered the citation, whether you appear in the body text or source cards, and which page on your site is being cited.
• If you do not appear: note that too. You will use this later.

While you are looking at the AIO, note which other businesses or sources appear, both in the body text and in the source cards. This is not competitive intelligence at scale. It is a directional signal about which entities Google currently considers authoritative for this query.

If your competitors are consistently cited and you are not, the gap is real. The question is which component is causing it: crawlability, entity signals, citation quality, content structure, schema, or AIO-specific eligibility. That diagnosis is what a GEO Audit provides.

• Note which competitors or other sources appear in the AIO for this query.
• Check the source cards to see which specific pages are being cited.
• Record this alongside your own citation status. You will want this for comparison over time.

Run the same process for each query on your list. Record the results for each one: whether an AIO appeared, whether you were cited and where, which competitors appeared, and which pages were cited.

After running five queries, you have the beginning of a picture. Not a complete picture. A single session is a snapshot, and AIO results vary by phrasing, location, and session. But a snapshot is more than you had before.

The query did not trigger an AIO. This query type may not be AIO-eligible. Review your query list and replace commercial or transactional queries with informational and research-oriented ones. The AIO check is only useful on queries that generate AIOs.

An AIO appeared but your business was not cited. You have a gap. The question is where it sits: is it a crawlability issue, an entity signal problem, a citation quality gap, a content structure issue, or an eligibility signal? Identifying which requires a structured assessment, not another manual check. What is GEO? covers the components. The GEO Audit covers how to diagnose which ones apply to your specific situation.

You appear in source cards but not in the answer body. Your content is being accessed but your brand is not being surfaced as the answer. This often points to content structure — answer capsule format, heading specificity, or the directness of the answer provided. How to Structure Content so AI Systems Actually Cite It covers the structuring principles that affect this.

You are named in the answer body. Record it. Note the query, the page cited, and the date. Keep this as a baseline for future checks. Appearances can disappear. An update to the AIO system, a change to your site, or stronger competition can shift who gets cited. Documenting where you are now gives you something to compare against later.

• Record your results in a simple format: query, AIO appeared (yes/no), cited in body (yes/no), cited in source cards (yes/no), competitors cited, page cited.
• If you found gaps: identify which of the four outcomes above applies to each query and use it to direct the next step.
• If you found citations: treat this as a baseline, not a destination. Note the query and page, and plan to recheck in four to six weeks.

A single clean session check tells you where things stand right now. It does not tell you whether you are improving, declining, or holding steady. AIO appearances are not fixed positions. The same query, run next week, may return a different AIO citing different sources. A source cited today may not be cited in a month.

For that kind of directional picture, you need a structured monitoring approach: a defined set of priority queries, run on a consistent cadence, with results recorded over time. That is the difference between knowing your position on one day and understanding whether your GEO work is actually moving the needle.

How to Read Your GEO Performance covers what systematic GEO monitoring looks like and what the results mean.

Why doesn’t Google show AIO appearance data in Search Console?

Google Search Console does not currently separate AIO citation data from organic click data in a way that answers this question directly. AIO citations may contribute to impressions and clicks recorded in GSC, but there is no dedicated AIO report. This is a known limitation of current tooling. Manual clean session testing is the method practitioners use in the interim.

How often should I run this check?

For a one-off orientation, once is useful. For ongoing monitoring, a monthly check across a defined set of 20-30 priority queries gives you a trend line. Running checks more frequently without a documented query set and methodology produces noise rather than signal. You are observing normal variation, not meaningful change.

What does it mean if the AIO gives wrong information about my business?

It means your entity signals are inconsistent or incomplete. AI systems fill gaps from whatever they can find. If the available source material is inaccurate, outdated, or contradictory, the generated response may be too. The fix is better source material: accurate, consistent signals across your site and external sources. This is entity optimisation and citation work, not a content correction.

Is appearing in source cards good or bad?

It is progress. Source card appearance means your content is being accessed and used as an AIO source. Not being named in the answer body is a content structure constraint in most cases. The content is relevant but not formatted for clean extraction. Improving content structure, particularly answer capsule format, is the most direct response.

If you want to move beyond one-off checks and understand what is actually driving — or limiting — your AI visibility, get in touch or talk to JB.

TL;DR

• There is no Google dashboard for AIO appearances. Manual clean session testing is the current best method.
• AIOs appear most consistently on informational and research queries. Test with the right query types or you will not see AIOs at all.
• Open a fresh incognito window, do not sign in, and run your queries on the correct country domain.
• Incognito is not perfectly clean. Location still affects results. Test from the location your business serves where possible.
• Look for your business in two places: the AIO body text (being named in the answer) and the source cards (being used as a source). They are different levels of citation.
• Note competitor citations and which pages are being cited. Record everything.
• Four possible outcomes: no AIO (wrong query type), AIO but not cited (a gap to diagnose), cited in source cards only (content structure may be the constraint), named in the answer (record as a baseline).
• One check is a snapshot. Systematic monitoring requires a defined query set run on a consistent cadence over time.