Research

We publish what we measure, including the results that go against us. Every study below carries its method, its limits and a downloadable dataset.

Every figure on one page

The citable numbers from all three studies, each with its denominator, its collection date and a permanent link.

AI local search statistics

First study

11.94%

of the businesses Google Maps returns get named by any AI assistant

AI Local Visibility Index

We asked ChatGPT, Claude, Gemini and Google AI Overviews for local recommendations across 48 US markets, then checked their answers against the businesses Google Maps actually returns for the same query.

4,657 businesses · 1,440 AI answers · collected 2026-08-15/16

Read the study

Second study

70.3%

of the businesses AI named were named by only one of the four assistants

The Consensus Gap

The same dataset, a second question: do the four assistants agree with each other? They mostly do not, which means being recommended by AI is not one outcome but four separate ones.

36 of 48 markets had no unanimous pick

Read the study

Third study

0.53%

of Google’s measured demand is what AI volume tools report for the same local queries

The Local Blind Spot

This time the subject is not AI but the tools that claim to measure it. We sent the same local queries to an AI volume endpoint and to Google’s own advertising data. Our own product had this wrong too.

96 queries · 12 categories × 8 US cities

Read the study

Browse the first study by city

The same 48 markets as pages you can read one at a time, one per city and one per city and category. Each lists the businesses every engine actually named, so you can check a market you know instead of trusting the average.

AI visibility by city

How we work

Measurement rules are written down and frozen before data is collected.
Query lists are generated from templates, not hand-picked, so nobody has to trust our choice of examples.
A missing measurement is never reported as a zero. The two are different claims and we keep them apart.
Every study publishes its raw data and the script that reproduces its figures.
Results that go against us get published too. The third study found a mistake in our own product and says so.

Find out where your business sits

Run the same questions against your own category and city.