ChatGPT named the same brand in every city; the one that searched the web named local firms

Short answer: ask ChatGPT for the best recruitment agency in your city, then ask the same question for another city in the same country. It is often the same name. For the same trade, the business it names most appears in 65% of a country's cities; for the only assistant that went and searched the web, 28%.

You want to know whether the AI recommends your business. The question assumes it looks at your city. For some trades it does not: it recites a national roster, the same from one city to the next. Ask for the best recruitment agency in London, then Manchester, then Leeds, and on down the list: ChatGPT answered Hays in all thirteen UK cities we asked about. One assistant was the exception, and it is the one that searched the web before answering.

For some trades the AI is not answering about your city. It is reciting a national roster.

What was measured

Waseit's eight studies ask, in every city and every trade, the question a customer would type: "what are the best {trade} in {city}?". For each country, trade and assistant we looked at the business named in the most of that country's cities, then measured in what SHARE of cities it reappeared. A genuinely local business can only be named where it operates; a national brand, a franchise or a platform can reappear everywhere. The measure covers 8 countries and 89 cities, captured between 31 July and 16 October 2026, each time in the language of the country. To compare the three assistants we restricted it to France, the United Kingdom and Brazil, the only countries where all three answered, and to trades covering at least ten cities: 49 lists per assistant.

Reading names out of an answer has a trap. Automatic extraction can take a subheading ("why it is recommended") or a piece of advice for a business name, and we keep only the first eight businesses in each answer. So we set aside a closed list of these false names, read one by one, and we quote only verified businesses: Hays, Robert Half and Century 21 are real brands, not reading artefacts.

Why does the same name come back in every city?

Because some of the assistants answer from memory. Queried without searching, ChatGPT and Gemini return, for certain trades, a learned roster - the big national networks - and serve it again whatever the city. We had already shown that half the recruiters the AI names are national networks and that for an accountant the AI repeats national firms, for a hotel never. The missing assistant was the one that searches the web: Perplexity, which read pages before answering, does not recite that roster. The business it names most reappears in only 28% of cities, and its answers often name neighbourhood firms.

In the trades where the roster is national, the same name comes back literally everywhere. Here are businesses ChatGPT named in every one of a country's cities that we asked about.

What it means if you are the business

If you run an independent agency - recruitment, estate, accounting, air-conditioning, a driving school - and you check once for "the best {your trade} in {your city}", you may read a list that has nothing to do with your city: the same national networks that come back everywhere. It is not that you are ranked low; it is that, for this trade, the assistant is not looking at your city. So the right question is not "am I first?" but "is the assistant answering with a national roster, or with businesses from my city?" - and, if it searches the web, "do the pages it reads mention me?". That is where your local presence matters again.

And your business - what do the assistants say?

That is what Waseit's free audit measures: it puts to ChatGPT and Gemini the eight questions your customers actually type, in your language and for your city, plus an awareness test that separates "the AI does not know you" from "the AI knows you and does not recommend you". In about 60 seconds you get a score out of 100, the competitors named in your place - including these big networks that come back everywhere - the questions where you do not appear, and which of the assistants actually searched the web rather than answered from memory. Free, no account and no card; an email address only unlocks the competitors' names and the action plan. It is a dated snapshot, not a ranking, and we promise no position.

What this does not prove

• Reading names out of an answer carries a risk: extraction can take a subheading or a piece of advice for a business, and we see only the first eight names in each answer. We set aside a closed list of these false names and quote only verified businesses, but the exact share of noise that remains, we do not know.
• A national brand reappearing in every city is not an error: a big recruiter does have offices everywhere. We are not saying these names are bad. We are saying that a local business, in these trades, does not enter the answer.
• Perplexity answered only in France, the United Kingdom and Brazil: the comparison between the three assistants is limited to those countries. And that it searched the web goes together with its more local answers, without our having shown that it is the cause.
• A single capture, between 31 July and 16 October 2026, through the assistants' programming interfaces, which approximate the consumer apps without reproducing them. The answers change from one day to the next; this is only a photograph of them.
• We measured a share of cities, not the quality of the lists. Nothing here says the named businesses are the best in the city, nor that the ones missing would be any better.