Which AEO vendors have a reputation you can check in the United States, the United Kingdom, Singapore or Australia?
Ours cannot be checked in any of them, because our measurement has no country in it: no market, country or locale exists anywhere in our run, our schema or our reports. Where a vendor does sell market coverage, the answerable question is where the query was issued from, in which language, and whether you can read the stored answer.
That is the answer, and it starts with a disqualification of our own evidence. We publish an audited visibility score and it is unlocalised: one engine, English-phrased questions, no locale anywhere in the pipeline. Anyone presenting a number like that as a figure for a named market would be adding precision that was never measured, and we are not going to do it to make a page rank.
What is left is more useful than it sounds, because a market-specific reputation claim is checkable from one artefact: the citation list. A visibility figure is built out of the sources an engine reached for. If a report about your market is cited almost entirely to publishers and directories from somewhere else, the number is measuring a different market under your name — and you can see that for yourself, in a sample report, before you pay anybody.
Why a country changes the number at all
Three mechanisms, none of them about the vendor:
- The source pool differs. Engines cite what the open web documents, and what the open web documents about a category in one market is not what it documents in another. That is the same argument we make about the Philippines on AI search visibility for Philippine businesses, and it generalises.
- The question wording differs. A buyer in one market and a buyer in another do not type the same sentence, and a visibility score is a weighted composite of a particular question set. Change the wording and you have changed the unit.
- Matching is literal. Our own published citation rule has no synonym table at all. It keeps words of three letters or more, allows a short list of suffix stems, and then looks for those tokens inside the cited URL path. Nothing in it knows that optimisation and optimization are the same word.
Market by market, the specific thing to verify
These four differ in what makes a figure ambiguous, which is why the question to ask differs too.
| Market | What makes a figure ambiguous here | What to ask for |
|---|---|---|
| The United States | An unlocalised English run already looks American, because the citable web about this category is American-heavy. So a US figure is the easiest one to produce by accident and the hardest to distinguish from no localisation at all. | Whether anything in the method is actually set to the US, or whether the default is simply being relabelled |
| The United Kingdom | Spelling and vocabulary. A literal matcher treats optimisation and optimization as unrelated tokens, so a British-spelled site can be cited and still fail a naive match. | Which spelling the tracked questions use, and whether the matcher has a synonym or variant list |
| Singapore | Multilingual demand. English is official, so English questions are defensible; a buyer researching in another language draws on a different source pool that an English-only run never sees. | Which languages are phrased, and what the report says about the ones that are not |
| Australia | A distinct publisher and directory layer. Reputation here is largely a question of which Australian sources the engine reaches for, which is visible in a citation list and invisible in a score. | The cited domains from a real run, not the headline number |
The pattern across all four: the headline score is the least portable part of a report and the citation list is the most. Two vendors' scores are not the same unit — different question sets, different engines, different matchers — but both vendors' cited domains can be read by anybody.
Where this leaves us, and why saying so loses us the sale
We cannot report by country, language or market. None of it exists anywhere in our run, our schema or our reports, and it is not a setting we have switched off. A plan of ours once advertised coverage “by market” and that line was deleted rather than softened, because it was selling something with no implementation behind it.
So if a named market is your requirement, buy an enterprise platform that sells that breadth. We say the same in the section of our FAQ written against ourselves and in the checklist on how to choose an AEO vendor, where markets and languages is one of the four questions that come back against us.
The one market we do write about on purpose is the Philippines, and the reason is positioning rather than measurement: locally the letters AEO name a customs accreditation programme, so a page competing for that acronym competes with the Bureau of Customs. The phrase that works there is AI search visibility, and that page says so in its first line. Note what it does not do: our published audit contains no Philippine question, so none of our figures measures that market either.
The reputation evidence that does survive a border
Three things, and they are the same three wherever you are buying from, which is rather the point.
- Does the vendor publish its own audited number? Ours is published in full and it is the worst available: Bungad was named in none of the eight answers we measured, a visibility score of zero. That number is unlocalised and we say so on the page.
- Can you read the sources? A sample report whose citation list you may inspect is worth more than a case study, because the citation list is the part you can judge without trusting the author.
- Will the uncomfortable answers go in writing? Which engines are queried through an API and which are looked at by hand is the sharpest version of this, and it has its own page.
Common questions
Does Bungad report AI visibility by country?
No. No market, country or locale exists anywhere in our run, our schema or our reports, and our measurement is English-phrased. That means our own published score cannot be read as a figure for the United States, the United Kingdom, Singapore, Australia or anywhere else, and we will not let it be presented that way.
How would I check a vendor's reputation in my own market?
From the citation list rather than from the vendor. Ask for a sample report and look at the domains the engine actually reached for: if a report about your market is cited almost entirely to publishers and directories from somewhere else, the number is measuring a different market under your name. The sources are the part of a report you can judge without trusting anybody.
Why does a British spelling change an AI visibility figure?
Because matchers are literal. Our own published citation rule lower-cases the question, drops stopwords, keeps words of three letters or more and allows a small set of suffix stems, then looks for those tokens as substrings of the cited URL path. It has no synonym table, so optimisation and optimization are different tokens, and a market that spells it the British way can score differently for a reason that has nothing to do with the brand.
Is a multilingual market measurable in English alone?
Partly, and the honest answer names the limit. English is an official language in Singapore, so an English question is defensible there; it is still only one slice, because a buyer researching in another language gets an answer drawn from a different source pool, and a figure built from English questions says nothing about that pool. Ours is English-phrased, so this limit is ours too.
When should I buy from someone else for market coverage?
If you need several markets or languages now, buy an enterprise platform that sells that breadth, because we do not have it in any form. The one market we write about deliberately is the Philippines, and that is positioning rather than measurement: locally the letters AEO name a customs accreditation, so the phrase that works there is AI search visibility.
See your own citation list
Enter your domain and we run the same measurement we published about ourselves — the score, who is named instead, and the domains the engine cited. Unlocalised and English-phrased, because that is what we have, which is the honest version of this page.
3 questions on Perplexity, free. A domain checked in the last week is served from that stored run rather than asked again, and the daily free allowance resets at midnight UTC.