Four external certifications, verifiable in one click
All four scores are measured by independent tools (Google, Mozilla, Scott Helme, Qualys). Click any card to re-run the test live.
Six commercial queries — five #1 positions (FR, EN, DE) and one #3 + Google AI Mode
Positions observed on Google.lu and Google.de depending on query language, Luxembourg geolocation, signed-out browsing. May 2026.
Six #1 positions on Google — and 12 documented AI citations below. Want to see how your site scores?
Test my site in 30 seconds →12 verifiable citations, six engines, three languages
Each citation is dated, attributed to a specific engine, and includes the original query. French and German verbatim quotes are provided in their source language, then translated into English.
How to read these verbatims: AI engines paraphrase from the pages ranking highest on the query. When one cites heliorank.lu, it may reuse phrasing from the content itself — so read these citations as « proof that the content is findable and exploitable », not as a human endorsement. The robust proof remains the classic organic position shown under each citation.
AEO is augmented technical SEO — not a separate channel
AEO is technical SEO with an extra markup layer — not a separate channel.
- →AI engines mainly cite pages ranking 1 to 3 on Google. A page buried on page 4 won't be paraphrased.
- →On 8 of the 12 citations above, heliorank.lu also ranks #1 to #3 organic on the same query.
- →The AEO delta = structured layer (Schema, llms.txt, FAQ, Organization graph). Good SEO makes content exploitable by LLMs.
Consistent citation across the Mozilla / public-scores vertical
heliorank.lu is also cited by Google AI Overview on four related queries beyond AEO. Topical authority signal on the web-security & public-scores vertical — Luxembourg geolocation, May 2, 2026.
- audit mozilla luxembourg
- expert mozilla luxembourg
- optimisation mozilla luxembourg
- expert optimisation mozilla luxembourg
Eleven captures — eight distinct queries (including one AI Overview citing Heliorank three times) plus two stability proofs (6 days)
Original dated captures in Brave private browsing (Luxembourg geolocation). Click any card to open the 1920px version. AVIF format (WebP fallback) lazy-loaded — negligible PageSpeed impact.
WattDevis spontaneously cited, Sitigo well understood and valued by ChatGPT, Gemini & Google AI
Sections 1-6 prove the method applied to heliorank.lu. Here is how it manifests on two other sites I personally operate — spontaneous citation for WattDevis (less saturated solar vertical), and fine understanding + explicit valuation for Sitigo as soon as you ask its opinion (Sitigo in launch phase facing Wix, WordPress, Shopify, Squarespace). This Sitigo valuation is no accident: it's the result of a 72-hour GEO sprint (May 17-19, 2026) that eliminated a major interpretation bug — before, Gemini was inventing a recurring subscription for Sitigo; after, Gemini correctly cites « one-time 69 € payment, no subscription ». Method: near-duplicate pruning (350 → 151 URLs), 60,000 original words written across 104 trade-city pages, SERP title refactor on 124 pages, 2 dedicated comparison pages, Schema Offer enriched with hasMerchantReturnPolicy. Detailed public case studies: WattDevis and Sitigo. Sites concerned: wattdevis.fr (solar quote comparator) and sitigo.fr (multilingual SaaS for artisan site generation). So this is the same method applied by the same operator across 3 distinct sites — not yet a commercial client case study. Original dated conversations, Brave private browsing except one logged-in capture.
How these citations were tested
Test protocol
Each query is run in a signed-out private session, Luxembourg geolocation. Screenshots are dated and stored with the exact query URL.
Verbatim sources
FR (Aperçu IA, ChatGPT FR) and DE (Übersicht mit KI) verbatim are reproduced in their source language and translated into English. No paraphrasing is applied — verbatim copy from the captures.
Non-determinism
AI answer engines and Google AI Overviews vary from one session to another based on geolocation, personalization, model version, and algorithmic updates. A citation present today may disappear tomorrow. The scoreboard presents a snapshot of a given period — not a guarantee of permanence.
What these signals actually measure
The HELIORANK methodology optimizes the technical signals leveraged by AI engines and classic SEO in parallel. These signals facilitate citation without guaranteeing it. A Lighthouse 100/100 score, a rich Schema graph (Organization, FAQPage, ItemList), a structured llms.txt, and a good organic SEO ranking statistically increase the probability of appearing in LLM responses — never a contractual promise.
Case study honesty
All the proofs above apply to 3 sites I personally operate: heliorank.lu, wattdevis.fr and sitigo.fr. So this is a methodological repeatability proof (same operator, 3 different verticals). Two structured case studies document the method: WattDevis (SEO + technical security, 2026) on the WordPress → static HTML migration and the Mozilla D → A+ shift in 30 days; and the Sitigo GEO sprint (May 17-19, 2026) on the complete 72-hour recovery (288 unindexed pages → 0, Gemini moving from a wrong interpretation to a correct citation of the 69 € one-time pricing). Two distinct verticals — SEO+security on one side, GEO/AEO on the other — same reproducible method.
Original captures
Dated captures are available on request during client audits or initial discovery calls. Public diffusion is deliberately limited to respect the terms of use of the engines concerned.
Frequently asked questions about this scoreboard
Are the AI citations in this scoreboard verifiable?
Yes. Every technical score (Lighthouse, Mozilla Observatory, SecurityHeaders, SSL Labs) links to a public tool you can re-run yourself on heliorank.lu. The AI answer-engine citations are dated and shown with the exact query; timestamped screenshots are kept and shared on request during audits.
What is the difference between AEO and GEO?
AEO (Answer Engine Optimization) optimises a site's presence in direct engine answers: Google AI Overview, ChatGPT, Perplexity. GEO (Generative Engine Optimization) covers more broadly how generative models understand and rephrase a brand. In both cases the foundation stays the same technical SEO: semantic structure, Schema.org, performance and llms.txt.
Does an AI citation today guarantee a citation tomorrow?
No. AI answer engines vary from one session to the next depending on geolocation, personalisation and model version. The scoreboard is a dated snapshot, not a guarantee of permanence. The methodology optimises the technical signals that statistically increase the probability of being cited, never a contractual promise.
Can this scoreboard be reproduced on your own site?
Yes. The HELIORANK AEO Setup implements the signals measured here: a structured llms.txt, an enriched Schema.org graph, Person and Organization nodes, then monthly monitoring of citations across ChatGPT, Gemini, Claude and Perplexity. The approach is documented in the two WattDevis and Sitigo case studies.
