We audited 550 companies across 22 cities against 94 checks that determine whether an AI engine can read, understand and cite a website. 370 of them (67.3%) score an F. The mean is 36.9 out of 100. Not one company scored an A.
550 companies · 22 cities · 94 checks · open dataset, CC BY 4.0
550
Companies audited
67.3%
Score an F
36.9
Mean score
24%
Score 50 or above
Failure is the norm, not the exception.
370 of 550 companies (67.3%) score an F. Only 132 companies (24%) reach 50 out of 100. The single highest score in the entire corpus is 75. These are established, well-funded companies, and almost none of them are set up to be cited.
The gap is structured data, and it is not close.
Across the 425-company cohort where category scores are directly comparable, structured data averages 15.3 out of 100 while crawler access averages 46.9. Most sites are reachable by AI crawlers and then give them nothing machine-readable to work with.
One missing markup type accounts for more failures than anything else.
Product/Service Schema is flagged as a critical gap on 288 companies, 52.4% of the corpus. The next two, llms.txt presence at 48.9% and reviews & ratings at 36.4%, are the same story: cheap, well-documented fixes that almost nobody has made.
Geography barely matters. Effort does.
New York leads the city table on 48.9 and Los Angeles trails on 14, a spread of 34.9 points. The spread inside a single city is routinely larger than the spread between cities, so this is not a market-maturity story. It is a per-company one.
Average readiness score across the 550 companies audited. Each city links to its full ranking.
| # | City | Companies | Average | Top scorer |
|---|---|---|---|---|
| 1 | New York | 25 | 48.9 | Datadog (65) |
| 2 | Toronto | 25 | 46.2 | Achievers (61) |
| 3 | Stockholm | 25 | 45.8 | Funnel (67) |
| 4 | Melbourne | 25 | 45 | Marketplacer (66) |
| 5 | Chicago | 25 | 44.8 | Sprout Social (73) |
| 6 | Paris | 25 | 43.4 | Spendesk (70) |
| 7 | Sydney | 25 | 43.2 | Raiz Invest (58) |
| 8 | Mumbai | 25 | 42 | Upstox (63) |
| 9 | Singapore | 25 | 41.4 | Syfe (56) |
| 10 | Amsterdam | 25 | 40.5 | Finom (58) |
| 11 | Sao Paulo | 25 | 39.2 | Cora (55) |
| 12 | Miami | 25 | 39 | Brikl (67) |
| 13 | Zurich | 25 | 37.9 | Scandit (61) |
| 14 | Bangalore | 25 | 37.3 | Jupiter (59) |
| 15 | Berlin | 25 | 36.8 | Babbel (60) |
| 16 | Sydney & Melbourne | 25 | 36.7 | Employment Hero (63) |
| 17 | Seoul | 25 | 32.8 | SOCAR (57) |
| 18 | Nairobi | 25 | 31 | Sun King (44) |
| 19 | Silicon Valley | 25 | 26.9 | Stripe (53) |
| 20 | Tokyo | 25 | 22.8 | Josys (75) |
| 21 | Tel Aviv | 25 | 16.1 | JFrog (57) |
| 22 | Los Angeles | 25 | 14 | ServiceTitan (56) |
Share of the 550 audited companies where each item was flagged as a critical gap.
Category averages across the 425 companies in 17 cities whose sub-scores share a common scale. See the methodology note below.
Structured data
15.3
out of 100
Content quality
27.2
out of 100
Trust and authority
29.3
out of 100
Technical delivery
40.6
out of 100
Crawler access
46.9
out of 100
F
370 companies
D
157 companies
C
19 companies
B
2 companies
C+
1 company
B-
1 company
No company in the corpus scored an A.
Cohort. 550 companies across 22 cities, selected as recognisable businesses headquartered in or strongly associated with each city, spanning consumer, B2B software, financial services and marketplaces. Scans ran between 2026-04-18 and 2026-06-17.
Measurement.Each company's public website was audited against 94 checks across 8 categories covering AI crawler access, structured data, content quality, trust and authority signals, and technical delivery. The composite is normalised to 0-100 and graded. Full methodology at /methodology.
What this does not measure. This is a readiness audit. It measures how legible a site is to an AI engine, not how often the brand is currently named in AI answers. Those are different measurements and we do not treat one as a proxy for the other.
A caveat we are not hiding. Category sub-scores changed scale between scanner generations: 17 of the 22 cities recorded them as percentages, the rest as raw point counts. Averaging the two together would produce a meaningless number, so every category-level figure on this page covers the 425-company percentage cohort only. Composite scores and grades are consistent across both generations and therefore span all 550 companies.
Published under CC BY 4.0. Journalists and analysts are welcome to quote any figure with attribution.
BrandCited. “State of AI Visibility.” 550 companies across 22 cities audited against 94 AI-visibility checks, 2026-04-18 to 2026-06-17. https://www.brandcited.ai/state-of-ai-visibility-2026
The same 94-check audit behind this study runs on any domain in about two minutes.
Scan your site free