Methodology
A useful score, not an empty promise.
The AI Readiness Score measures observable signals on your website. It does not claim to reproduce the private ranking or retrieval systems used by Google, OpenAI, Anthropic, Perplexity or any other provider.
Our principles
- Every score must be explainable.
- Every failed check must include evidence.
- Every recommendation must be actionable.
- Passing the checks cannot guarantee an AI mention or citation.
- Results are a diagnostic snapshot and may change as the website changes.
Category weighting
| Category | Points |
|---|---|
| Crawlability and indexingWhether crawlers can reach your pages and find your sitemap. | 20 |
| Business identity and structured dataWhether machines can identify who you are and what you offer. | 20 |
| Content clarity and answerabilityWhether your page text is readable, titled and clearly structured. | 20 |
| Trust and authority signalsWhether ordinary trust pages and details are present and linked. | 15 |
| External presence and consistencyHow your business appears across sources beyond your website. | 15 |
| llms.txt and machine-readable contentOne small component: a plain-text summary file for AI clients. | 5 |
| Rendering and performanceWhether content exists in the HTML rather than only after scripts run. | 5 |
What the free test retrieves
The free test requests four public addresses: your submitted homepage, /robots.txt, /sitemap.xml and /llms.txt. A limited number of redirects is followed. Responses are read as text with a size limit and a strict timeout. Scripts from the scanned site are never executed, and no cookies or credentials are sent.
Crawler access is reported as a matter of fact. Blocking an AI crawler is often a deliberate policy decision, so it is never scored as an error.
How the score is calculated
Each check has a fixed number of points. Where a check cannot be completed because no evidence is available, it is marked “Not tested” and excluded from the calculation rather than guessed. The displayed score is renormalised across the completed checks only, and the audit coverage figure tells you what proportion of the full 100 points was actually assessed.
In this first version, external presence and loading performance are always outside the free homepage scan. That is why coverage is below 100% even for a website that passes everything.
Score bands
- 85 to 100 — Excellent foundation
- 70 to 84 — Good, with opportunities
- 50 to 69 — Needs improvement
- 0 to 49 — Significant gaps
Quick Test and Complete Website Audit
Quick Test
Free and open to everyone. It reads your homepage plus robots.txt, sitemap.xml and llms.txt.
- One page of content, plus your machine-readable files
- Good for a first diagnosis and quick wins
- Cannot judge consistency across your other pages
Complete Website Audit
Requires a free account. It examines up to 25 selected public pages and looks at the site as a whole.
- Site-wide consistency of titles, descriptions, identity and structured data
- Internal architecture, broken links and sitemap agreement
- A prioritised implementation plan and a downloadable PDF report
The two scores are calculated with the same weighting but from different evidence, so they are labelled separately and should not be compared directly. Neither test measures whether AI assistants actually mention or cite your business. That is AI Presence monitoring, and it is not part of this version.
AI Readiness and AI Presence are different things
AI Readiness is whether your website is technically and editorially prepared to be understood. AI Presence is whether your business is actually mentioned or cited in live AI-generated answers. This version measures readiness only. Presence monitoring will follow as a paid feature.
On llms.txt specifically: it is a small, optional convenience file worth five of the hundred points. Its presence is not a guaranteed method of appearing in AI results, and we do not award significant credit merely for having one.