Study 02 · Supporting Market Volume · Great Britain

llms.txt Adoption in Great Britain

A descriptive market breakdown. These figures show what Great Britain looked like; what the study demonstrates — that llms.txt adoption does not predict AI-crawler openness — is established in the pooled cross-market benchmark, where the sample supports inference.

Research instrumentPTODA C01 Crawler v1.3
Scan date17 June 2026
Sample754 domains · 603 scannable · 75 adopters
12.4%
of scannable Great Britain businesses have adopted llms.txt

75 of 603 scannable domains publish llms.txt, of which 84.0% are valid files. Great Britain has the lowest conformance of the four markets and a higher infrastructure non-response rate, which narrows the paired sample.

The market figures

Adoption, conformance, contradiction, and the reconfirmed access benchmark for Great Britain — all drawn from the pooled study tables.

12.4%
llms.txt adoption
75 of 603 scannable domains.
84.0%
Conformance of adopters
63 of 75 published a valid file.
32.4%
Contradiction (observed)
24 of 75 adopters also block AI crawlers.
38.8%
Access benchmark
207 / 533 domains with an observable policy block at least one AI retrieval crawler. Reconfirms the frozen v1.2 figure.
75
Adopters observed
The llms.txt adopter count for Great Britain — the paired-analysis base for this market.
603
Scannable domains
Domains where the relevant layer could be observed (denominator).

Interpretation note

Descriptive, not inferential. This market-level report is descriptive. The number of llms.txt adopters observed within a single market is insufficient to support independent statistical inference regarding the relationship between llms.txt adoption and AI-crawler access. The inferential finding for this study is reported in the pooled four-market benchmark, which combines all observable domains into a single paired analysis (n = 1,617).

This separation is deliberate. A single market's adopter sample is too small for its contradiction rate to be read as a proven market effect — Great Britain’s 32.4% is observed, not demonstrated. For the hypothesis test, the null result, the effect size, and the interpretation, see the pooled AI Accessibility & Machine-Readable Guidance Benchmark, which carries all inferential claims for this study.