The AEO score (0–100) is a weighted average of six category scores. Each category score is the share of available points earned across its checks. Checks that can't be evaluated (for example, no images on the page, or AI analysis unavailable) are excluded instead of counted as failures.
Overall = Σ (category score × category weight) ÷ 100. Grades: A ≥ 85, B ≥ 70, C ≥ 55, D ≥ 40, F < 40.
Measured vs estimated
- Measured: observed directly from your page HTML, HTTP responses, robots.txt, XML sitemap, llms.txt, Google PageSpeed Insights (Lighthouse), RDAP domain registration records, Wikidata and OpenStreetMap (Nominatim).
- Estimated: a structured AI judgment of your page text against the fixed rubrics below. The model returns a probability-weighted position on each rubric with a confidence value, which we show in the evidence column. Treat these as directional.
- Not measured: Google review counts, Business Profile completeness and backlink volume need paid APIs. They are covered in the paid detailed audit and never guessed.
What we don't claim
No tool can see inside ChatGPT, Gemini or Perplexity ranking systems. This audit measures publicly observable signals that the engines document or demonstrably use: crawler access, machine-readable entity data, answerable content, technical health and third-party corroboration. A high score raises your chances of being cited; it doesn't guarantee it. Live prompt testing is part of our detailed audit.
Recommendations
Every failing or weak check maps to a recommendation. Our AI analysis then rates how much each fix would improve AI visibility for your type of business (for example, opening hours matter more for a café than for a SaaS company). Fixes rated not relevant are dropped, and the rest are ordered by that rating, weighted towards quick wins.
Categories and checks
AI Search Readiness
20% of overallCan AI crawlers reach, read and reuse your content?
| Check | Max points | Source |
|---|---|---|
| AI answer-engine crawlers allowed | 25 | Measured |
| AI training crawlers allowed | 10 | Measured |
| Server responds to AI crawler user-agents | 15 | Measured |
| Content visible without JavaScript | 20 | Measured |
| llms.txt file | 5 | Measured |
| No noindex / nosnippet restrictions | 10 | Measured |
| Quotable, citable statements | 15 | Estimated (AI) |
Content & Answerability
20% of overallDoes your content answer real questions clearly enough to be cited?
| Check | Max points | Source |
|---|---|---|
| Content depth | 10 | Measured |
| Heading structure | 10 | Measured |
| Question-style headings | 10 | Measured |
| FAQ section | 8 | Measured |
| Lists and tables | 5 | Measured |
| Readability | 7 | Measured |
| Visible freshness signals | 5 | Measured |
| Directly answers customer questions | 15 | Estimated (AI) |
| Specific, verifiable facts | 10 | Estimated (AI) |
| Experience & expertise signals | 10 | Estimated (AI) |
Structured Data
18% of overallIs your business described in machine-readable schema.org markup?
| Check | Max points | Source |
|---|---|---|
| JSON-LD structured data present | 15 | Measured |
| Organization / LocalBusiness schema | 20 | Measured |
| Business schema completeness | 20 | Measured |
| WebSite schema | 5 | Measured |
| FAQ / HowTo schema | 10 | Measured |
| Service / Product / Article schema | 10 | Measured |
| BreadcrumbList schema | 5 | Measured |
| Review / rating schema | 5 | Measured |
| Structured data parses correctly | 10 | Measured |
Technical SEO
16% of overallIndexability, metadata, performance and mobile basics.
| Check | Max points | Source |
|---|---|---|
| HTTPS | 10 | Measured |
| Homepage loads (HTTP 200) | 10 | Measured |
| Title tag | 8 | Measured |
| Meta description | 8 | Measured |
| Canonical URL | 6 | Measured |
| Mobile viewport | 6 | Measured |
| Language declared | 4 | Measured |
| XML sitemap | 8 | Measured |
| robots.txt | 5 | Measured |
| Open Graph tags | 5 | Measured |
| Image alt text | 5 | Measured |
| Mobile performance (PageSpeed) | 15 | Measured |
| Server response time | 5 | Measured |
Business Profile
13% of overallAre your name, address, phone and offering unambiguous and consistent?
| Check | Max points | Source |
|---|---|---|
| Business name identifiable | 5 | Measured |
| Phone number visible | 12 | Measured |
| Email or contact form | 6 | Measured |
| Address or service area | 12 | Measured |
| Contact page | 6 | Measured |
| About page | 6 | Measured |
| Google Maps link or embed | 6 | Measured |
| Listed on OpenStreetMap | 10 | Measured |
| Name/address/phone consistency | 10 | Measured |
| Clear who / what / where | 15 | Estimated (AI) |
| Pricing information | 6 | Estimated (AI) |
| Google rating & review volume | – | Not measured |
Brand Authority
13% of overallDo independent public sources confirm your brand exists and is trusted?
| Check | Max points | Source |
|---|---|---|
| Domain age | 15 | Measured |
| Wikidata entity linked to your site | 15 | Measured |
| Wikipedia article | 10 | Measured |
| Official social / directory profiles linked | 15 | Measured |
| Testimonials / reviews on site | 10 | Estimated (AI) |
| Privacy policy & terms | 10 | Measured |
| Named authors / people | 5 | Measured |
| Awards, certifications, press, partners | 10 | Estimated (AI) |
| Backlinks & brand mentions | – | Not measured |
AI crawlers we test
Answer-engine / live search: OAI-SearchBot, ChatGPT-User, PerplexityBot, Claude-SearchBot, Bingbot (Copilot), Googlebot (AI Overviews). Model training: GPTBot, Google-Extended, ClaudeBot, CCBot, Applebot-Extended. robots.txt is evaluated per RFC 9309 (most specific user-agent group, longest matching rule). We also request your page using GPTBot and PerplexityBot user-agent strings to detect firewall or CDN blocks.
Estimated-check rubrics
- Quotable statements (0–4): from “no quotable statements” to “concise definitive answers with specifics throughout”.
- Directly answers customer questions (0–4): covers what, for whom, where, cost and how to start.
- Specific, verifiable facts (0–3) and experience & expertise signals (0–3).
- Entity clarity (0–3): can the business be summarised accurately in one sentence?
- Yes/no probabilities: pricing shown, testimonials shown, third-party validation mentioned.
Questions about the method? Contact us.
