Dashboard
AI Labs Audit
Last audit: 05/15/2026 12:39 PM
AGS Methodology v2 active — calibrated by 5 LLM judges, honest metrics (unbiased Share of Voice, spontaneous mentions, Wilson confidence intervals).
AI Visibility
C+
59/100 +3
95% CI: 74 – 75 · on 630 responses
Technical GEO
B
71/100
AEO Readiness
C+
64/100
Global diagnostic
Your AI visibility score at a glance.
Key Metrics
View
Visibility Score
58%
58.7%
Mention Rate
54%
54.6%
AI Share of Voice
9%
9.6%
Average Position
83%
1.0
Sentiment
58%
5.8/10
Topic Coverage
83%
83.4%
AI presence
How AI engines cite your brand — when, why, against whom.
Evolution Over Time
Technical Metrics
Qualite des mentions
No data — run a new audit
Brand Safety
23 factual error(s)
121 response(s) where the AI doesn't truly know the brand
[claude-haiku-4.5] Most facts about the company's founding date, founder, and features are based on self-reported claims; some details are uncertain or unverified.
[gpt-5-mini] The information is factually accurate regarding the company's website and available public data, but lacks independent validation.
Brand Safety v2 (LLM-judge): 99% safe responses
Performance by AI engine
Which model cites you best: ChatGPT, Sonar, Claude, others.
Sector Radar
Your AI footprint compared to your sector
Services numériques
AI Labs Audit
Previous audit
Native vs Web Score
62/100
Score
Score without Internet
What AIs know about you by defaultVS
Web +13.3
75/100
Score
Score with Internet
What AIs find when searching
Web search noticeably improves your scores. Keep enriching your online content while working on your presence in AI training data.
AI Model Breakdown
Global view all languages
| AI Model | Score | Sentiment | Position | Mention Rate |
|---|---|---|---|---|
|
Gpt 5 Mini:Online
WEB
|
6.7/10 |
6.8
|
#1.5 |
92.9%
|
|
Gpt 5 Mini
NATIF
|
6.1/10 |
5.7
|
#1.7 |
87.1%
|
|
Sonar
WEB
|
6.2/10 |
6.5
|
#1.7 |
87.1%
|
|
Claude Haiku 4.5
NATIF
|
3.4/10 |
2.5
|
#2.7 |
85.7%
|
|
Grok 4.1 Fast:Online
WEB
|
6.6/10 |
6.4
|
#1.6 |
84.3%
|
|
Gemini 2.0 Flash 001:Online
WEB
|
5.6/10 |
5.3
|
#2.3 |
80.0%
|
|
Grok 4.1 Fast
NATIF
|
5.4/10 |
5.3
|
#1.6 |
78.6%
|
|
Gemini 2.0 Flash
NATIF
|
5.0/10 |
4.2
|
#2.0 |
77.1%
|
|
Claude Haiku 4.5:Online
NATIF + WEB
|
4.3/10 |
4.3
|
#1.7 |
60.0%
|
GEO Section Analysis
Detailed scores by prompt category
S1 — Visibility & Awareness
4/4 categories tested
5.3/10
Brand awareness
4.4/10
5.3
56%
-2.1
Topical coverage
4.3/10
6.9
44%
+0.8
Digital footprint
6.1/10
6.4
100%
+0.8
Multilingual visibility
6.5/10
7.1
94%
+0.3
S2 — Authority & Trust (E-E-A-T)
5/5 categories tested
5.0/10
E-E-A-T
5.6/10
5.1
94%
+0.4
Citations & Mentions
3.9/10
4.3
89%
+0.5
Expertise & Know-how
5.1/10
5.6
94%
+0.3
Content authority
5.7/10
6.4
89%
Social signals
4.5/10
5.9
94%
+0.7
S3 — Sentiment & Perception
3/3 categories tested
5.3/10
General sentiment
4.9/10
5.1
100%
+1.0
Reputation & Image
4.7/10
5.0
100%
+0.5
Crisis management
6.2/10
5.9
100%
+0.2
S4 — Competitive Positioning
5/5 categories tested
6.2/10
Direct comparison
6.2/10
7.0
89%
-0.2
Competitive duel
6.6/10
6.9
94%
Top N ranking
5.1/10
7.0
65%
-1.0
Alternatives & Substitutes
6.0/10
5.9
89%
-0.6
Differentiation
7.3/10
7.4
100%
+0.8
S5 — Search Intent
8/8 categories tested
5.9/10
Direct / Navigational
6.0/10
6.6
89%
-0.7
Informational
6.9/10
8.0
89%
-0.4
Comparative
5.2/10
6.3
89%
-0.4
Transactional
6.5/10
6.7
100%
-0.4
Targeted intent
4.5/10
7.2
50%
-1.4
Local & Geographic
5.3/10
5.9
78%
-1.4
Voice & Conversational
6.3/10
6.2
100%
-0.6
Problem solving
6.4/10
6.8
100%
S6 — Customer Journey & Recommendation
5/5 categories tested
5.5/10
AI recommendation
5.5/10
6.5
82%
Persona targeting
6.1/10
6.4
94%
-0.3
Customer micro-moments
3.3/10
8.2
22%
+0.5
B2B business matching
6.2/10
6.0
89%
Advanced purchase intents
6.6/10
6.8
100%
GEO technical audit
Is your site crawlable, citable, optimised for AI?
Technical Metrics
SSR / AI Crawlability
97
SSR: 90%
Robots.txt: 100%
Acces IA: 100%
llms.txt: 100%
Framework: Next.js
Entity Health Check
50
Wikidata: 0
Properties: 0
Schema.org: 100
sameAs: 83
Citation Readiness
49
Sources: 80
Stats: 0
Experts: 0
Factual dens.: 30
Paragraphs: 22
H2 Questions: 25
Direct answers: 100
Source Authority
Owned: 637
Third-party: 1787
Your pages most cited by AIs. Optimize them to amplify your voice.
Your amplifier pages
111×
87×
45×
39×
37×
AI Perception
visibility 1685x
monitoring 737x
audit 687x
visibilité 661x
content 644x
audits 620x
tools 591x
pages 505x
mentions 502x
optimization 481x
Based on 630 AI responses
Advanced signals 2026
30 signals
AI Visibility Files (AEO)
robots.txt
(15%)
100%
llms.txt
(5%)
100%
Schema.org
(15%)
82%
sitemap.xml
(10%)
100%
AQA Standard
(10%)
70%
OpenGraph
(5%)
100%
RSS feed
(5%)
0%
dashboard.aeo_breakdown_quotable_blocks
(15%)
0%
dashboard.aeo_breakdown_freshness_90d
(10%)
100%
Answer density (Q/A)
(5%)
0%
Entity consistency
(5%)
0%
C+
Grade
robots.txt
Present
28.3 KB
AI Bots :
- GPTBot
- OAI-SearchBot
- ChatGPT-User
- ClaudeBot
- Claude-SearchBot
- Claude-Web
- anthropic-ai
- PerplexityBot
- Perplexity-User
- Google-Extended
- GoogleOther
- Applebot-Extended
- meta-externalagent
- Meta-ExternalAgent
- Bytespider
- CCBot
- Bingbot
- Amazonbot
- cohere-ai
sitemap.xml
Present
889 URLs
Sitemap index (5 child sitemaps)
llms.txt
Present
8.8 KB
llms-full.txt
Present
72.2 KB
Scan vue bot IA — 10 pages analysees avec User-Agent GPTBot
(home + 3 FAQ, 2 blog, 2 guide, 2 random)
AQA
Standard
3 Q&A enriched
— trouve sur 7/10 pages
3 dated
3 sourced
3 changelog
To reach FULL :
- author credentials (jobTitle, affiliation) on each question
Schema FAQ
Present
trouve sur 10/10 pages
FAQPage: 7 questions
HowTo: 4 steps
Article
BlogPosting
BreadcrumbList
FAQPage
HowTo
Organization
OpenGraph
5/5
trouve sur 10/10 pages
— 100% de couverture complete (>=4/5 tags)
+3 bonus (dates, author)
RSS Feed
Absent
No RSS or Atom feed detected.
No <link rel="alternate"> RSS/Atom on 10 pages.
Complete GEO Checklist
Optimization for generative AI engines
Contenu recupere via : Vue Chrome (utilisateur)
|
Fiabilite 100%
|
81ms
GEO Checklist
(15%)
98%
SSR Crawlability
(25%)
97%
Entity Health
(25%)
50%
Citation Readiness
(20%)
49%
Mention/Citation
(15%)
67%
B
Grade
AI Crawlability
The robots.txt file controls which bots can crawl your site. If it blocks AI crawlers (GPTBot, ClaudeBot...), your content will never be indexed by generative AI engines.
AI Bots :
GPTBot
OAI-SearchBot
ChatGPT-User
ClaudeBot
Claude-SearchBot
Claude-Web
anthropic-ai
PerplexityBot
Perplexity-User
Google-Extended
GoogleOther
Applebot-Extended
meta-externalagent
Meta-ExternalAgent
Bytespider
CCBot
Bingbot
Amazonbot
cohere-ai
Actions to complete
- Verify that robots.txt is accessible at /robots.txt (HTTP 200)
- Verify no global block User-agent: * / Disallow: /
- Verify GPTBot is NOT in a Disallow directive
- Verify ClaudeBot is NOT in a Disallow directive
- Verify Claude-SearchBot is allowed
- Verify OAI-SearchBot is allowed
- Verify PerplexityBot is allowed
- Verify Google-Extended is allowed
- Verify Applebot-Extended is not blocked
- Ensure public content directories (/blog/, /services/, /about/) are accessible
How to verify
Run an AI Labs Audit before/after changes. Check server logs for AI bot requests within 2-4 weeks. Test manually with curl -A "GPTBot" https://yoursite.com/.
Run an AI Labs Audit before/after changes. Check server logs for AI bot requests within 2-4 weeks. Test manually with curl -A "GPTBot" https://yoursite.com/.
The sitemap helps AI crawlers discover all your important pages. Without a valid sitemap, some pages may never be indexed by AI engines.
889 URLs
— Index (5 sitemaps)
Actions to complete
- Verify /sitemap.xml returns HTTP 200
- Verify sitemap is referenced in robots.txt via Sitemap: https://yoursite.com/sitemap.xml
- Verify the XML is valid (no parsing errors)
- Verify <lastmod> tags are present on each URL
- Verify <lastmod> dates match actual update dates
- Verify there are no error URLs (404, 500) in the sitemap
How to verify
Check Google Search Console for sitemap indexation status. Monitor server logs for AI bot sitemap access.
Check Google Search Console for sitemap indexation status. Monitor server logs for AI bot sitemap access.
The llms.txt file is an emerging standard that helps AI quickly understand your site. It provides a structured summary of your business and key pages.
8.8 KB
Actions to complete
- Create a /llms.txt file at the site root
- Add an H1 title with the company name
- Add a summary blockquote (2-3 sentences describing the business)
- List key pages with URL and short description (home, services, about, pricing, FAQ, blog)
- Use valid Markdown format
- Test accessibility with curl https://yoursite.com/llms.txt
- Optional: also create a /llms-full.txt file with full content
How to verify
Monitor server logs for llms.txt file access. Note: direct impact is not yet proven, it's an AI readiness signal.
Monitor server logs for llms.txt file access. Note: direct impact is not yet proven, it's an AI readiness signal.
HTTPS is a trust prerequisite for all AI engines. A site without HTTPS is considered unreliable and will rarely be cited as a source.
Actions to complete
- Verify the site is accessible via HTTPS
- Verify the SSL certificate is valid and not expired
- Verify HTTP → HTTPS redirect is in place (301 permanent, NOT 302)
- Verify no mixed content (HTTP resources on HTTPS pages)
- Verify all internal URLs use HTTPS
How to verify
Immediate test via browser or curl -I https://yoursite.com. Certificate must be valid.
Immediate test via browser or curl -I https://yoursite.com. Certificate must be valid.
An overly restrictive WAF (Web Application Firewall) can block AI crawlers like GPTBot or ClaudeBot, preventing AI engines from indexing your content.
Bot IA accepté
— HTTP 200
No known WAF/CDN detected
Server:
nginx
Actions to complete
- Test site access with GPTBot user-agent: curl -A "GPTBot/1.0" → should return 200
- Test with ClaudeBot: curl -A "ClaudeBot/1.0" → should return 200
- Test with PerplexityBot: curl -A "PerplexityBot" → should return 200
- If Cloudflare: verify "Bot Fight Mode" is not in aggressive mode
- If Cloudflare: verify AI Crawl Control has desired AI crawlers set to "Allow"
- Check custom WAF rules: no rule should generically block user-agents containing "bot"
- Verify no CAPTCHA challenge is returned to AI bots
- Verify WAF rules and AI Crawl Control settings do not contradict each other
How to verify
Re-run curl tests after changes. Monitor Cloudflare analytics for AI bot requests: they should show as "Allowed".
Re-run curl tests after changes. Monitor Cloudflare analytics for AI bot requests: they should show as "Allowed".
Structured Data
Schema.org structured data helps AI understand your page content and context. It's a strong signal to be cited as a reliable source.
Types:
Article
BlogPosting
BreadcrumbList
FAQPage
HowTo
Actions to complete
- Add Organization JSON-LD script on homepage (name, url, logo, description, contactPoint, sameAs)
- Add WebSite JSON-LD script on homepage
- Add Article or BlogPosting JSON-LD script on each blog article
- Add FAQPage JSON-LD script on pages with FAQs
- Add BreadcrumbList JSON-LD script for breadcrumb navigation
- Add Person JSON-LD script on author/team pages
- For service/product pages: add Product or Service with Offer
- Validate each schema with Google Rich Results Test
- Verify JSON-LD is server-side rendered (SSR), not injected via client JavaScript
How to verify
Validate via Google tools. Run an AI Labs Audit and compare the Structured Data score.
Validate via Google tools. Run an AI Labs Audit and compare the Structured Data score.
Open Graph tags improve how AI and social networks understand your pages. They provide structured title, description, and image.
5/5 tags requis
Actions to complete
- Verify og:title is present on all pages (unique per page)
- Verify og:description is present (complementary to title)
- Verify og:image with a quality image (1200x630px recommended)
- Verify og:image:alt with descriptive text
- Verify og:url points to the canonical URL
- Verify og:type is present (website for home, article for articles)
- For articles: add article:published_time and article:modified_time
- Also add Twitter Card tags (twitter:card, twitter:title, twitter:description, twitter:image)
- Test rendering with Facebook Sharing Debugger and LinkedIn Post Inspector
How to verify
Test sharing a URL on LinkedIn/Facebook/Twitter and verify the preview is correct and complete.
Test sharing a URL on LinkedIn/Facebook/Twitter and verify the preview is correct and complete.
The canonical tag prevents duplicate content and helps AI crawlers index the correct version of each page.
Actions to complete
- Verify each page has a <link rel="canonical" href="..."> tag
- Verify canonical is self-referential (points to the same page)
- Verify consistency between canonical, og:url and actual URL
- Verify canonicals don't point to 404s or redirects
- Verify pages with UTM parameters have a canonical to the clean version
- Verify protocol consistency (all HTTPS)
How to verify
Crawl the site with Screaming Frog or similar tool and check the "Canonical" column for inconsistencies.
Crawl the site with Screaming Frog or similar tool and check the "Canonical" column for inconsistencies.
AI engines prioritize recent content. The dateModified property in JSON-LD schemas signals that your content is up to date.
Actions to complete
- Add "dateModified" to all Article/BlogPosting JSON-LD scripts
- Add "datePublished" if missing
- Verify dateModified is more recent than datePublished
- Display the update date visibly on the page ("Updated on...")
- Update the corresponding <lastmod> in sitemap.xml
- Set up a regular update process (at least quarterly)
How to verify
Validate JSON-LD with Rich Results Test. Run AI Labs Audit and verify the check turns green.
Validate JSON-LD with Rich Results Test. Run AI Labs Audit and verify the check turns green.
Content Quality
The title tag is the first signal AI reads to understand a page's topic. A unique, descriptive title of proper length maximizes your chances of being cited.
Actions to complete
- Each page has a unique <title>
- Length between 30 and 60 characters
- Contains the main keyword or page topic
- No generic title ("Home", "Page", domain name alone)
- Consistent with the page H1
- No duplication with other site pages
- Includes brand name (at end of title, after separator)
How to verify
Crawl the site and extract all titles. Verify uniqueness and length. Track CTR evolution in Google Search Console.
Crawl the site and extract all titles. Verify uniqueness and length. Track CTR evolution in Google Search Console.
The meta description provides a structured summary to AI. An informative, unique description per page increases your chances of being selected as a source.
Actions to complete
- Each page has a meta description
- Length between 120 and 160 characters
- Informative and descriptive content (not hollow marketing)
- Unique per page
- Contains a direct answer or actionable summary
- Includes an implicit call to action when relevant
How to verify
Crawl the site to check coverage. Track CTR in Google Search Console.
Crawl the site to check coverage. Track CTR in Google Search Console.
Structured Q&A is the preferred format for AI to extract information. A well-structured FAQ with FAQPage schema maximizes your chances of being cited.
Actions to complete
- Identify the 5-10 most frequent customer/prospect questions
- Write direct, factual answers in 2-4 sentences each
- Integrate the FAQ section on key pages (services, pricing, home)
- Use <h2> or <h3> tags for each question
- Implement FAQPage schema in JSON-LD with question/answer pairs
- Validate schema via Rich Results Test
- Write questions in natural conversational language
How to verify
Validate FAQPage schema. Test questions in ChatGPT, Claude and Perplexity to see if the site appears in responses.
Validate FAQPage schema. Test questions in ChatGPT, Claude and Perplexity to see if the site appears in responses.
A logical heading hierarchy helps AI understand your content structure. Each H2 section should be self-contained and extractable by an LLM.
H1:1
H2:10
H3:32
Actions to complete
- Each page has a single H1 describing the main topic
- H2s structure the main sections
- H3s detail the subsections
- No level skipping (no H1 → H4 directly)
- Headings are descriptive and self-explanatory
- No headings used for styling purposes
- Each H2 section is relatively self-contained (extractable alone by an LLM)
How to verify
Use an SEO audit tool (Screaming Frog, Semrush) to validate heading structure.
Use an SEO audit tool (Screaming Frog, Semrush) to validate heading structure.
Pages with fewer than 2000 words are rarely cited by AI. Substantial content with data and concrete examples increases your credibility.
5863 words
Actions to complete
- Check word count of main content on each page (excluding nav, footer)
- Pages under 2000 words should be enriched or consolidated
- No "under construction" or placeholder pages
- Service pages should have at least 500 words of substantial content
- Include data, concrete examples, and use cases
- Avoid generic or duplicated content from other sources
How to verify
Count words on key pages. Run audit again to verify the check turns green.
Count words on key pages. Run audit again to verify the check turns green.
A regular blog with expert content increases your authority in the eyes of AI. Long, specialized articles are most likely to be cited.
Actions to complete
- Create a Blog or News section on the site
- Publish a first in-depth article on your domain of expertise
- Each article must have: named author, visible publication date, Article/BlogPosting schema, 800+ words, images with alt text
- Set a publication rhythm (minimum 1 article/month)
- Cover customer questions (guide, comparison, how-to formats)
- Set up internal links between articles and service pages
- Ensure articles are indexed in the sitemap
How to verify
Track organic traffic to the blog section. Regularly test covered topics in ChatGPT/Claude/Perplexity to see if the site is cited.
Track organic traffic to the blog section. Regularly test covered topics in ChatGPT/Claude/Perplexity to see if the site is cited.
LLMs (AI Overviews, Perplexity, Claude) prioritize pages containing blockquotes, definition lists (<dl>), data tables, callouts (stat/tldr/fact boxes) and ordered lists. Without these formats, your page is less citable with attribution.
Actions to complete
- Convert key stats into <blockquote> with source
- Add a <dl> 'Key terms' with 5-10 definitions
- Replace ranking <ul> with numbered <ol>
- Wrap key figures in <aside class='stat'> or <div class='callout'>
- Add at least 1 comparative <table> (>=2 data rows)
How to verify
Score 0-100 based on count of 5 sub-types on homepage. Pass if score >= 30. See Citation-Readiness card in dashboard.
Score 0-100 based on count of 5 sub-types on homepage. Pass if score >= 30. See Citation-Readiness card in dashboard.
LLMs prioritize extracting short paragraphs (40-50 words) that directly answer an implicit question. Text composed only of long descriptive paragraphs isn't 'answer-ready'. Target: 30%+ paragraphs in answer format.
Actions to complete
- Identify key questions your clients ask
- Reformulate long paragraphs into 'implicit question + 40-50-word answer' blocks
- Place a standalone answer at the start of each h2/h3 section
- Avoid paragraphs >100 words without syntactic break
How to verify
Automatic check: ratio of 40-50-word paragraphs over total. Pass if >=30%.
Automatic check: ratio of 40-50-word paragraphs over total. Pass if >=30%.
LLMs build a semantic representation of your brand from the 30 most co-occurring words in your pages. If these terms lack sufficient density, LLMs do not "know" your brand on these themes. Princeton KDD 2024.
Actions to complete
- Identify 30 key terms in your domain (competitor analysis)
- Check density (>=0.5%) of each term on homepage and key pages
- Enrich content to fill underdensified terms
- Align glossary / FAQ with these 30 terms
How to verify
Automatic check: average density of top-30 co-occurring terms vs target. Pass if average >=0.5%.
Automatic check: average density of top-30 co-occurring terms vs target. Pass if average >=0.5%.
Authority & Trust
A detailed About page strengthens your credibility (E-E-A-T). AI uses it to understand who you are and decide whether to cite you as a reliable source.
Actions to complete
- Page is accessible from main navigation or footer
- Contains company history (founding date, founders)
- Contains verifiable legal information (registration number, address, legal form)
- Describes mission, values, positioning
- Mentions certifications, labels, partnerships
- Contains at least 400 words of substantial content
- Organization schema implemented with the same information
- Team or office photos (proof of real existence)
How to verify
Ask ChatGPT, Claude and Perplexity "What is [company name]?" and verify if returned information matches the About page.
Ask ChatGPT, Claude and Perplexity "What is [company name]?" and verify if returned information matches the About page.
AI values content written by identifiable experts with qualifications. Detailed bios strengthen E-E-A-T signals.
Actions to complete
- Each content author has a biography on the site
- Bio contains: full name, position, expertise area, career summary
- Mentions certifications, degrees, or years of experience
- Professional photo
- Link to LinkedIn profile
- Person schema implemented (name, jobTitle, description, sameAs, worksFor)
- Blog articles display author name and bio
- No content published under "Admin" or "The team" without attribution
How to verify
Verify Person schema is valid. Test "[expert name] + [field]" in AI to see if expertise is recognized.
Verify Person schema is valid. Test "[expert name] + [field]" in AI to see if expertise is recognized.
Social profiles strengthen trust signals for AI. They confirm the company's real existence and enrich the Knowledge Graph.
twitter
youtube
facebook
linkedin
Actions to complete
- Social media links present on the site (footer or contact page)
- All links are functional (no 404s or deleted profiles)
- Profiles are active (last publication < 3 months)
- Brand name is consistent between site and all profiles
- Organization schema with sameAs property listing all profiles
- Minimum 2 active networks (LinkedIn is most important for B2B)
- Profiles link back to the website
How to verify
Check consistency of Google Knowledge Panel for the brand. Validate Organization schema.
Check consistency of Google Knowledge Panel for the brand. Validate Organization schema.
The E-Commerce Directive 2000/31/EC requires displaying the registration number (SIREN/SIRET in France, Company Number in UK, KVK in Netherlands, IDE in Switzerland, etc.) in the legal notices. AIs use this information to validate the existence and legitimacy of the business — a major E-E-A-T signal.
Actions to complete
- Identify your registration number for your country (SIREN/SIRET, Company Number, KVK, IDE, BCE, etc.)
- Add it to your Legal Notice / Impressum / Terms page
- For France: add SIREN, SIRET, RCS and registration city
- For UK: add Company Number (8 characters) and place of registration (Companies House)
- For Germany/Austria: add Handelsregisternummer (HRA/HRB + number) in the Impressum
- For Switzerland: add IDE number (CHE-xxx.xxx.xxx)
- Ensure the number is in plain text (not in an image)
How to verify
Check that the registration number is visible as text in the legal notices. The automated audit detects it via scraping.
Check that the registration number is visible as text in the legal notices. The automated audit detects it via scraping.
An email and phone number directly accessible on the site are mandatory (E-Commerce Directive). AIs value easily contactable businesses as more reliable and worthy of recommendation.
Actions to complete
- Add a contact email in the footer or contact page (mailto: format preferred)
- Add a phone number (tel: format preferred)
- Ensure these are in plain text, not only behind a contact form
- Ideally: add Schema.org ContactPoint markup
How to verify
Check for mailto: and/or tel: links on the site. The automated audit detects them.
Check for mailto: and/or tel: links on the site. The automated audit detects them.
The director, manager or publication director name in the legal notices is a strong transparency signal. AIs associate content with an identifiable person to evaluate Expertise and Authority (E-E-A-T).
Actions to complete
- Add the full name of the director/manager in the legal notices
- Specify their role (CEO, Manager, Director of Publication)
- Ideally: add Schema.org Person markup with sameAs to LinkedIn
- If applicable: mention director qualifications or certifications
How to verify
Check that the director name appears in the legal notices or impressum. The automated audit detects it.
Check that the director name appears in the legal notices or impressum. The automated audit detects it.
If your brand appears under multiple variants (e.g., 'CITEL', 'Citel SA', 'CITEL France', 'Citel-Group'), LLMs don't merge these entities and your citations are diluted. Target: <=2 distinct variants across the entire site.
Actions to complete
- Choose a unique canonical form of the name (e.g., "CITEL")
- Audit the site (header, footer, og:site_name, Schema Organization) to remove variants
- Align Schema.org Organization.name with canonical form
- Check consistency on LinkedIn, Wikipedia, third-party profiles
How to verify
Automatic check: variant detection on homepage + key pages. Pass if <=2 variants.
Automatic check: variant detection on homepage + key pages. Pass if <=2 variants.
Freshness & Signals
The <meta name="robots" content="noai"> tag blocks AI from using your content. If present, your site is invisible to generative AI engines.
Actions to complete
- Check for absence of <meta name="robots" content="noai"> in <head>
- Check for absence of X-Robots-Tag: noai HTTP header
- If present, remove it to allow AI indexation
How to verify
The automated audit detects the noai meta tag and X-Robots-Tag header.
The automated audit detects the noai meta tag and X-Robots-Tag header.
An RSS or Atom feed signals to AI crawlers that your site publishes regularly. AI engines use feeds to discover fresh content. Content less than 30 days old receives 3.2x more citations.
Actions to complete
- Add a <link rel="alternate" type="application/rss+xml"> in <head>
- Ensure the feed is accessible (HTTP 200, valid XML)
- Include at least the 10 most recent modified articles/pages
- Verify each item has a <pubDate> or <updated> date
How to verify
The AEO section of the dashboard automatically detects RSS or Atom feeds.
The AEO section of the dashboard automatically detects RSS or Atom feeds.
AQA is an open standard that enriches your FAQs with AI-ready metadata: per-question dates, per-answer sources, revision history. AI systems prioritize content with verifiable provenance.
Actions to complete
- Implement AQA Basic markup on your FAQ pages (dates + sources)
- Add dateCreated and dateModified on each question
- Add a citation (source URL) on each answer
- Declare conformanceLevel in the Article block
- Validate with the AQA validator: github.com/sarsator/aqa-specification
How to verify
The AEO section of the dashboard automatically detects AQA presence and level (Basic/Standard/Full).
The AEO section of the dashboard automatically detects AQA presence and level (Basic/Standard/Full).
Content updated within the last 90 days is prioritized by AI. Content less than 30 days old receives 3.2x more citations. After 13 weeks without updates, content loses citation eligibility.
Actions to complete
- Verify dateModified is present in Schema.org JSON-LD
- Update key pages at least every 90 days
- Add current statistics (current year) in content
- Maintain a blog with regular publications
How to verify
The Citation Readiness v2 module automatically evaluates content freshness.
The Citation Readiness v2 module automatically evaluates content freshness.
Pages updated within the last 90 days are cited 3.2x more by LLMs (Princeton KDD 2024). Check dateModified (Schema.org) or HTTP last-modified header. If main page >90d, plan a visible update (recent figures, fresh sources).
Actions to complete
- Check dateModified on key pages via Schema Article/WebPage
- Update key content (stats, sources, dates) every <90 days
- Display the last-update date visibly (page footer)
- Add a changelog/revision history on important pages
How to verify
Automatic check: HTTP last-modified header + Schema dateModified. Pass if <=90 days on homepage.
Automatic check: HTTP last-modified header + Schema dateModified. Pass if <=90 days on homepage.
Sitemap & llms.txt Diagnostic
Sitemap.xml
Index (5 sitemaps)
via robots.txt
URLs
908
With lastmod
904
Avg. freshness
9 days
6 sitemaps déclarés dans robots.txt
Sub-sitemaps :
sitemap-pages.xml
160 URLs
sitemap-blog.xml
212 URLs
sitemap-glossaire.xml
509 URLs
sitemap-vitrine.xml
4 URLs
sitemap-videos.xml
4 URLs
908 URLs indexed
4 URLs without lastmod
Average age: 73.6 days
Page freshness 100 — click to expand
+ 70 more pages
llms.txt
+ llms-full.txt
Referenced pages
44
Tokens
1024
Title
AI Labs Audit
Extra pages (full)
+691 pages
Sitemap not in llms
875
44 pages referenced
1024 tokens
llms-full.txt: 691 additional pages
875 sitemap pages missing from llms.txt
Action plan
What you can do now, by priority.
Recommended Action Plan
12 actions
Action plan progress
0 Done
0
12
0%
6 high priority
5 medium priority
Other actions
12
Respond to negative reviews on public platforms
Respond professionally to negative reviews on Google, forums and review sites.
5.85/10
54.63%
Medium Priority
Sentiment
⚡ Quick fix
Medium Impact
+1.3 pts · 30min
Verify and fix sameAs links
Verify all sameAs links in Schema.org point to valid and active pages.
Medium Priority
GEO Optimization
⚡ Quick fix
Medium Impact
+1.0 pts · 30min
Add quotable blocks (callouts, definition lists, stat tables)
LLMs (AI Overviews, Perplexity, Claude) prioritize blockquotes, <dl> definition lists, data tables and callouts (class="callout", "tldr", "stat") for verbatim attributed citation. A page without any of these blocks is rarely quoted directly. Add 2-3 clear stat callouts + 1-2 structured definitions on your key pages.
Score : 0/100 — seuil 30
Détail : blockquotes: 0, definition lists: 0, data tables: 0, callouts: 0, ordered lists: 0
Medium Priority
GEO Optimization
🔧 Moderate
Medium Impact
+1.3 pts · 1.0h
Standardize the company name across all pages
LLMs identify entities through cross-page textual consistency. Using more than 2 variants of the name (e.g. "CITEL", "Citel SA", "Group VTS") muddies entity resolution. Pick ONE canonical name and use it everywhere, with variants only in sameAs tags and legal mentions.
Medium Priority
GEO Optimization
⚡ Quick fix
Medium Impact
+1.3 pts · 1.0h
Afficher le numéro d'immatriculation
Le SIREN/SIRET n'est pas visible. Les IA associent la transparence légale à la fiabilité.
Low Priority
Presence
⚡ Quick fix
Low Impact
+0.6 pts · 30min
Standalone 40-50 word answer after each subheading
For each H2/H3, the first sentence must be a complete standalone 40-50 word answer that can be extracted and cited as-is. Avoid "It depends…" or sentences that refer to prior context. Preferred format for AI Overviews and Perplexity direct citation.
High Priority
GEO Optimization
🔧 Moderate
High Impact
+1.5 pts · 1.5h
Add descriptive alt-text to images
Beyond SEO, multimodal LLMs read alt-texts to interpret images. Describe visual content and brand link, not just keywords.
Medium Priority
GEO Optimization
🔧 Moderate
Medium Impact
+0.8 pts · 1.0h
Launch a targeted positive content strategy
Create and distribute content highlighting strengths, customer testimonials, case studies and concrete results.
15.28%
54.63%
High Priority
Sentiment
🔧 Moderate
Medium Impact
+1.3 pts · 2.0h
Create or enrich the Wikipedia/Wikidata page
If the brand is notable enough, create a Wikipedia page with verifiable sources. Otherwise, enrich the Wikidata entry.
High Priority
Native Knowledge
🏗️ Project
High Impact
+2.3 pts · 5.0h
Identify and address sources of negative sentiment
Use audit source tracing to identify URLs that generate negative sentiment in AI responses.
83.37%
5.85/10
High Priority
Sentiment
🏗️ Project
High Impact
+1.9 pts · 5.0h
Fix contradictions detected across AI engines about your brand
When ChatGPT/Claude/Gemini/Perplexity give contradictory facts about your brand (50 vs 200 employees, 2015 vs 2018, Paris vs Lyon HQ), it signals training-source confusion. Identify the wrong sources and update them.
1.0
62.14
High Priority
Sentiment
🏗️ Project
High Impact
+1.5 pts · 4.0h
Publish and signal original data (studies, surveys, benchmarks)
LLMs prioritize sources that produce their own data. Publish 1-2 studies/surveys/benchmarks per year with explicit attribution ('Our research shows', 'According to our data'). +30-40% AI visibility boost (Princeton GEO).
High Priority
Native Knowledge
🏗️ Project
High Impact
+1.7 pts · 6.0h
Backlog (2)
Actions deprioritized by the GEO engine. Click to activate.
Verify and improve encyclopedic entries
+5.0 pts · 5.0h
· sentiment
Outreach to domains co-cited with your brand
+5.5 pts · 8.0h
· share_of_voice