AI Visibility Report
Latest edition: October 2026. We asked ChatGPT, Claude, Gemini and Grok the questions buyers ask, three times each, about 44 kinds of business in 108 places, and counted every name. These are the patterns across all of them. Claude was asked each question once on 1,755 of the 2,580 pages.
- Calls made
- 80,082
- Pages
- 2,580
- Pages ranked
- 2,580
- Assistants
- 4
Across the 2580 ranked pages of the October 2026 Index, 56% of the names an assistant gave came back when it was asked the same question again, and any two assistants shared 21% of their names. The average page gathered 29 different businesses, 45% of them named once, and its leader appeared in 59% of answers.
Findings
Averages over the ranked pages of the edition.
Asked the same question again, an assistant repeats 56% of its names.
For each assistant and question, the names that came back in at least two of three runs, as a share of every name those runs gave. The rest appeared once and did not return, which is why a single answer is a poor guide to what an assistant says about a business.
Any two assistants share 21% of the names they give.
For each pair of assistants that both named someone on a page, the names both gave as a share of the names either gave, averaged over pairs and pages. Being named by one assistant says little about the others.
Asked for a local business, Grok named no one in 76% of its answers.
An empty answer is not a failed call: the model replied and named nobody. Asked about software, Grok named no one in 10% of its answers. Asked for a local business, ChatGPT named no one in 6%, Claude named no one in 58% and Gemini named no one in 0%.
61% of the businesses in a ranking were named by only one assistant.
Of the businesses a page lists, those that a single assistant named in two or more answers and no other assistant named at all.
A page gathers 29 different names, and 45% of them are named only once.
Distinct businesses named across every answer on a page, and the share of them that a single answer named. Those one-off names are counted but never listed.
The most-named business appears in 59% of answers.
The leader's share on the average ranked page. On 1,446 of 2,580 ranked pages at least one business was named by every assistant asked there.
With web search on, 43% of the names were ones the model also gave without it.
From 3,260 answers with search switched on, compared with the same page's answers without it.
What changed from the previous edition
September 2026 compared with October 2026.
The categories, places, models, call routes and number of reruns can change between editions. These are differences between published samples, not a controlled measure of improvement in AI recommendations.
- Answers counted
- September 2026
- 87,618
- October 2026
- 80,059
- Difference
- -7,559
- Names repeated on a rerun
- September 2026
- 48.9%
- October 2026
- 55.7%
- Difference
- +6.8 percentage points
- Names shared by two assistants
- September 2026
- 20.2%
- October 2026
- 20.6%
- Difference
- +0.4 percentage points
- Ranked businesses named by one assistant only
- September 2026
- 55.9%
- October 2026
- 60.5%
- Difference
- +4.6 percentage points
The previous figures are preserved in the published recommendations study. Read the study
Download and cite the latest edition
This address always shows the latest published edition. Use its edition DOI when you need a citation that preserves the data you read.
- Latest edition
- October 2026
- Ranking data
- Download the data as CSV
- Edition DOI
- 10.5281/zenodo.23188442
- Archived files
- Download the archived files on Zenodo
- DOI for the latest archived edition
- 10.5281/zenodo.23082266
To cite this edition: AIGROW (2026). AI Visibility Index, October 2026 edition [Data set]. Zenodo. https://doi.org/10.5281/zenodo.23188442
Past editions
Each edition keeps its original dataset and DOI. The latest report stays here.
- September 2026
- 10.5281/zenodo.23082267Download CSV
By assistant
Every call, including the ones that failed. The last three columns are the answers that replied but named no business: about a local trade, about software and about a consumer brand.
By country
Questions were asked in the language of each country. Pages that were not ranked count here too.
| Country | Pages | Ranked | Names per page | Named again on a rerun |
|---|---|---|---|---|
| Italy | 194 | 194 | 28 | 50% |
| United States | 269 | 269 | 35 | 58% |
| United Kingdom | 194 | 194 | 30 | 57% |
| Ireland | 69 | 69 | 26 | 62% |
| Canada | 119 | 119 | 31 | 60% |
| Australia | 119 | 119 | 31 | 60% |
| Spain | 144 | 144 | 27 | 56% |
| Mexico | 94 | 94 | 25 | 56% |
| Germany | 194 | 194 | 31 | 50% |
| Austria | 94 | 94 | 28 | 52% |
| Switzerland | 169 | 169 | 30 | 51% |
| France | 194 | 194 | 28 | 53% |
| Belgium | 119 | 119 | 28 | 53% |
| Netherlands | 119 | 119 | 31 | 54% |
| Portugal | 69 | 69 | 27 | 59% |
| Brazil | 94 | 94 | 25 | 60% |
| Sweden | 94 | 94 | 26 | 60% |
| Denmark | 69 | 69 | 27 | 62% |
| Norway | 69 | 69 | 23 | 66% |
| Poland | 94 | 94 | 27 | 55% |
Rankings from the AI Visibility Index
See which businesses and products ChatGPT, Claude, Gemini and Grok named for the categories and places below.
Citing this report
AIGROW, AI Visibility Report, October 2026, aigrow.gg/research/ai-visibility-report. The figures and the data are published under CC BY 4.0: reuse them with a link to this page.
Where do you stand in these numbers?
The free scan asks the same assistants your own customers' questions and shows who they name instead of you.
Edition October 2026 · published by Guglielmo Paesano, founder of AIGROW · data under CC BY 4.0