Evidence
What 31 complete website audits revealed
These figures come from an authorised portfolio of business websites. The current sample is primarily WordPress and should not be treated as representative of the entire web.
No individual website, client or domain is ever identified here.
31
Complete website audits
65/100
Average AI Readiness Score (based on 31 websites)
60/100
Average mobile performance (based on 26 websites)
5 further audits finished only partly and are excluded from every figure on this page.
Score distribution
- 0 to 390
- 40 to 547
- 55 to 6910
- 70 to 8414
- 85 to 1000
Common findings
- 60% had missing or incomplete structured data (based on 25 websites).
- 56% had a sitemap problem (based on 27 websites).
- 96% had no useful llms.txt file (based on 27 websites).
- 92% had weak titles or descriptions (based on 25 websites).
- 100% did not pass: Privacy and terms pages are published (based on 27 websites)
- 100% did not pass: Links in llms.txt resolve (based on 3 websites)
- 96% did not pass: Each page has exactly one H1 (based on 25 websites)
- 92% did not pass: Meta descriptions are present and distinct (based on 25 websites)
- 78% did not pass: An about page exists and was readable (based on 27 websites)
- 74% did not pass: A contact page exists and was readable (based on 27 websites)
- 60% did not pass: Pages contain enough substance to be useful (based on 25 websites)
- 56% did not pass: Examined pages appear in the sitemap (based on 27 websites)
A percentage only counts websites where that check genuinely completed. A measurement that could not be taken is never counted as a failure.
Website platforms observed
- WordPress: 26 of 31 websites (84%)
- Unknown: 5 of 31 websites (16%)
- WordPress: average score 69/100 across 26 websites
Platform detection is based on observable signals in the homepage and its response headers. It never affects a score, and it is not a judgement of any platform.
How these figures are produced
- Only fully complete audits of authorised websites are counted. Partly finished audits are excluded.
- One result per website: the most recent complete audit is used.
- Findings are counted once per website, never once per page.
- A percentage uses only the websites where that check genuinely completed.
- These findings describe this audited sample. They are not an industry-wide study.
- Platform detection is recorded for context only and does not affect any score.
- The make-up of the portfolio can influence the averages shown here.
- No domains, page addresses, contact details or account information are published.
- Figures are recalculated as new complete audits are added, so they will change over time.
These statistics describe observable website signals only. They are not a measure of how often any business is mentioned by an AI assistant.