Skip to content
Waseit
← All articles
AEOGEO

Across 13 US cities, only 21 names travel

AI assistants name 2,599 different businesses across 13 cities. 88% of them appear in a single city, only 21 are cited in eight cities or more, and two in all thirteen. The national layer of the answers is thin; the rest is local.

2,599

different businesses named across 13 cities

September 20, 20267 min read
In this article
  1. What we measured
  2. The national layer fits on one line
  3. City by city, the field is local
  4. What this does not prove
  5. What this changes for you

Short answer: AI does not answer with a national directory, it answers with local names. Across 13 US cities and 15 trades, assistants name 2,599 different businesses. 88% of them appear in one city only. Just 21 names are cited in eight cities or more, and two in all thirteen: Robert Half and Aerotek, both staffing agencies.

88%

appear in a single city

21

names cited in eight cities or more

The competitor taking your place is almost never a national brand. It is a business from your own city that the AI learned somewhere.

What we measured

The data is the Waseit US study collected on 13 September 2026: 15 trades in 13 cities, two questions asked of ChatGPT and Gemini, 780 answers. We pooled every business named, then counted, for each name, how many cities it appears in.

A name that appears in three cities or more is what we have called, since 14 September, a name that travels. Directories are counted separately, and advice headings slipped into the lists are removed.

The national layer fits on one line

Of the 2,599 businesses named, 135 appear in at least three cities, that is 5%. 21 appear in eight cities or more. Two are cited in all thirteen.

Here are all twenty-one, by number of cities: Robert Half (13), Aerotek (13), Randstad USA (12), Kforce (12), Insight Global (12), A-1 Driving School (12), A-1 Locksmith (11), Roto-Rooter (10), Marsh & McLennan Agency (10), HUB International (10), TEKsystems (10), RSM US LLP (9), OfficeTeam (9), Lucas Group (9), Benjamin Franklin Plumbing (8), Brown & Brown Insurance (8), Gallagher (8), Mister Sparky (8), Pop-A-Lock (8), Midas (8), All Star Driving School (8).

Three trades supply almost all of that list: staffing, insurance broking and driving schools, joined by plumbing, locksmiths and auto repair through their franchise networks. In the other trades there is no national layer at all.

City by city, the field is local

Waseit US study, collected 13 September 2026. 60 answers per city (15 trades, 2 questions, ChatGPT and Gemini). "Citations to a name that travels": share of citations going to a name present in at least three cities for the same trade. "Names unique to the city": share of the businesses named there that are named nowhere else.
CityAnswers naming a businessBusinesses per answerCitations to a name that travelsNames unique to the city
Houston83%4.531%63%
Dallas88%4.829%66%
Chicago88%4.926%69%
Philadelphia88%4.925%67%
New York85%5.124%72%
Denver83%4.624%70%
Austin87%4.522%69%
Phoenix87%4.821%70%
Atlanta83%4.821%71%
Los Angeles93%5.320%68%
San Diego87%4.720%67%
Miami90%4.919%74%
Seattle93%5.215%77%

The shape of the answer barely changes from city to city: between 83% and 93% of answers name someone, and they cite between 4.5 and 5.3 businesses. What changes is the names. In every city, between 63% and 77% of the businesses named are named in no other city.

And in 49% of the answers that name someone, not one of the names appears in three cities. One answer in two is therefore entirely local.

What this does not prove

That a travelling name is one company.

"A-1 Locksmith" in eleven cities is almost certainly unrelated locksmiths sharing a name; Mister Sparky, Pop-A-Lock and Benjamin Franklin Plumbing are franchise networks, so one brand and local operators. We count NAMES, not companies.

That the national layer is thin everywhere.

It is real in staffing, insurance and driving schools; our 15 trades do not cover the rest.

A perfect count.

Two spellings of one business can count as two names, which inflates the single-city share; advice headings that escape the cleaning inflate it too.

Complete answers.

Our length cap (600 tokens for ChatGPT, 1,400 for Gemini) cuts part of the lists, and four answers per trade and city remain a sample.

What an app user sees

, nor a change over time: one collection, on 13 September 2026, through the APIs.

What this changes for you

You are not competing against a national brand.

In most trades the spot is taken by a business from your own city, not by a name you would recognise.

Look at the names in YOUR city.

Two thirds to three quarters of the businesses named there are named nowhere else: a national ranking tells you nothing about your market.

In staffing, insurance and driving schools

, the national layer is real: check who holds it in your city before concluding the field is open.

One appearance is not enough.

As the 17 September edition showed, what counts is coming back from one question and one engine to the next.

Our free audit asks ChatGPT and Gemini about your trade in your city, and names the businesses cited instead of you. They are almost always your neighbours.

Frequently asked questions

Do AI assistants mostly name national brands for local businesses?
No. Across 13 US cities and 15 trades they name 2,599 different businesses, 88% of which appear in a single city. Only 135 appear in at least three cities, and 21 in eight cities or more.
Which businesses are named in the most cities?
Robert Half and Aerotek, two staffing agencies, are named in all thirteen cities. Then come Randstad USA, Kforce and Insight Global in twelve, followed by A-1 Driving School, A-1 Locksmith, Roto-Rooter, Marsh & McLennan Agency, HUB International and TEKsystems.
Does the city change how AI answers?
Barely: between 83% and 93% of answers name a business, with 4.5 to 5.3 names per answer depending on the city. It is the names that change: 63% to 77% of the businesses named in one city are named in no other.
How many answers are these figures based on?
780 answers collected on 13 September 2026 through the ChatGPT and Gemini APIs, for 15 trades in 13 US cities, that is 60 answers per city.