AI Overview checker: which keywords trigger AI Overviews?
Paste your keywords and get an instant average AI Overview trigger likelihood. See which queries likely show a Google AI Overview, and how much of your traffic is exposed to click loss. A directional estimate from query patterns, using vendor-reported trigger rates this site has not been able to verify.
| Keyword | Query type | AIO likelihood | Est. CTR loss (pos. 1) |
|---|
Trigger rates by query type
Short answer: paste up to 20 keywords, one per line. The checker labels each one by query type (comparison, review, question, informational or commercial) and gives a rough likelihood that Google shows an AI Overview for it. It does not search Google. It reads the wording of your keywords and applies vendor-reported rates this site has not verified, so treat the output as a first sort, not a measurement.
AI Overviews don't appear evenly. They cluster on informational and comparison queries and stay away from transactional ones. This tool classifies each keyword by pattern and applies a likelihood weight per type. The weights reflect query-type trigger rates that circulate in vendor reporting. Their samples are undisclosed, and the AI Overview statistics hub grades the breakdown partial, not independently verified. Read the ordering, not the decimals:
| Query type | Pattern | Weight used (vendor-reported, unverified) |
|---|---|---|
| Comparison | "X vs Y", "versus" | ~95% |
| Review / Best | "best…", "top…", "review" | ~86% |
| Question | how / what / why / can / does… | ~86% |
| Informational | general, no buy intent | ~50% |
| Commercial / Transactional | buy / price / deal / near me | <10% |
The headline number is simply the average trigger likelihood across your keywords: not a visibility score, and not a measure of how often AI cites you. The estimated CTR loss on triggered queries uses the Ahrefs figure of ~58% (vendor-reported, position 1 only, informational queries, December 2025 data). It is a conditional effect on triggering queries, not a site-wide loss (where that number comes from, and who it actually describes). It's a directional estimate. For exact live AIO presence per keyword, pair it with a SERP tracker.
What the sample list returns
The box loads with six keywords so you can see the output before pasting your own. Here is what the checker does with them:
| Keyword | Query type | Likelihood |
|---|---|---|
| notion vs airtable | Comparison | 95% · High |
| best crm for agencies | Review / Best | 86% · High |
| how to get cited by chatgpt | Question | 86% · High |
| what is generative engine optimization | Question | 86% · High |
| buy running shoes | Commercial | 8% · Low |
| seo pricing | Commercial | 8% · Low |
The average comes out at 62, and four of the six are counted as likely to trigger an overview (anything at 50% or above). In plain terms: two-thirds of this list sits on the kind of query where AI Overviews tend to appear, so those are the pages to check first on the live results page.
How to use it on your own list
- Export your top keywords by clicks from Search Console or a rank tracker. Up to 20 at a time.
- Paste them in, one per line. Results update as you type.
- Sort your attention by the High rows. Search each one on Google, signed out, and note whether an overview appears and who it cites.
- For keywords where an overview shows and you are not cited, work on being the source it quotes.
What this checker can't tell you
- It never looks at Google. Every result is a guess from wording. Whether an overview actually shows depends on the query, country, device and date, and it changes.
- The rules are simple word matches, checked in order. Any keyword with "vs" or "versus" is a comparison; then "best", "top", "review", "cheapest" or "alternatives" make it a review query; then a question word (how, what, why, can, does, is and so on) makes it a question; then buying words (buy, price, cost, deal, near me and similar) make it commercial. So "where to buy running shoes" counts as a question, not commercial.
- English only. Keywords in other languages mostly fall through to Informational at 50%.
- The CTR column is one borrowed number. The ~58% figure is the Ahrefs estimate for position 1 on informational queries. It is shown for every likely-triggering row as context, not calculated for your keyword.
- It says nothing about citations. A high likelihood means an overview probably appears, not that you are or aren't in it.
Google itself says there is no special optimization needed to appear in AI Overviews, and that the usual snippet controls (nosnippet, max-snippet and so on) apply to them (Google Search Central). If you want to see what AI engines tend to cite, the most-cited domains in AI Overviews page and the citation readiness checker are the next steps.
You can't fix what you can't see
Most sites have no idea what share of their keywords now return an AI answer instead of a list of links. If much of your traffic sits on queries that show an overview, some of those clicks are likely being absorbed. The only durable fix is to become the source the AI Overview cites. This checker is step one: see the exposure, then price it and fix it.
Frequently asked questions
How does this estimate AI Overview exposure without a SERP API?
How do I check if a keyword actually shows an AI Overview?
What counts as "at risk"?
Do I need special optimization to appear in AI Overviews?
Can I stop my content appearing in AI Overviews?
nosnippet, data-nosnippet, max-snippet or noindex. These also limit how your page appears in regular results, so use them carefully.Does Search Console show AI Overview traffic?
Why does a keyword I know has an overview show Low?
My whole list is high-risk. Now what?
Related tools
llms.txt Generator
Generate and validate a spec-correct llms.txt for your site in one click.
Open →
Schema Markup Generator
FAQ, Article, HowTo & Person JSON-LD, ready to paste.
Open →
AI Traffic-Loss Calculator
Model how much organic revenue AI Overviews could cost you - and what to recover.
Open →One experiment. Every week.
The field notes in your inbox - one thing I tested, the raw numbers behind it, and what it means for getting cited by AI.