How AI reads this page
ChatGPT, Claude, Gemini and Perplexity don't look at colours, images or animations. They receive text, headings, links and structured data. Below is exactly what they read when they open “How We Generate Questions”.
The AI receives 369 words in 4 sections, 0 links and 2 blocks of structured data.
Text
The page content in Markdown, without menus or graphics. It is the format closest to how a language model reads text.
# How We Generate Questions > The quality of Geosnap's analysis depends on the quality and coverage of the generated questions. The system employs a structured approach that simulates URL: https://geosnap.ai/en/doc/how-questions-are-generated Language: English Version IT: https://geosnap.ai/doc/how-questions-are-generated Publisher: Geosnap (Maind Group S.r.l.), https://geosnap.ai ## The Generation Logic The quality of Geosnap's analysis depends on the quality and coverage of the generated questions. The system employs a structured approach that simulates how real users interact with AIs. Geosnap does not use static keywords. The system generates complete conversational questions, formulated as a real user would pose them to ChatGPT, Gemini, or Perplexity. This approach reflects how people actually use AIs: with natural language questions, not search strings. ## The Three Levels of Intent ### Informational Intent Questions oriented towards knowledge and education. The user wants to understand, learn, and delve deeper. Typical examples: - "What is [your product category]?" - "How does [your industry technology/approach] work?" - "What are the benefits of [solution]?" These questions measure if your brand is associated with authoritative content in your field. ### Commercial Intent Questions about evaluation, comparison, and selection. The user is considering different options. Typical examples: - "What are the best [product/service] for [use case]?" - "Comparison between [Solution A] and [Solution B]" - "[Category]: which one to choose in 2025?" These questions measure if your brand appears in AI recommendations when a user is deciding. ### Transactional Intent Questions with purchase, subscription, or direct action intent. Typical examples: - "How much does [product/service] cost?" - "How to buy/subscribe to [solution]?" - "Alternatives to [competitor] with a lower price?" These questions measure if your brand is present at the decision-making moment. ## Coverage and Variety Geosnap generates hundreds of variants for each intent, varying formulation, angle, and specificity. This ensures that results do not depend on a single formulation but represent a statistically reliable picture. Variety is important because LLMs are sensitive to formulation: the same question posed slightly differently can produce completely different answers. Covering many variants reduces noise and produces more robust data. ## Updating the Questions The questions are updated over time to reflect market evolution, changes in user language, and new trends in your sector. This ensures that the analysis remains relevant and accurate.
Structured data
Schema.org data (JSON-LD) embedded in the page code. It tells the AI explicitly what the page is about and who published it.
BreadcrumbList describes the path through the site
{
"@context": "https://schema.org",
"@type": "BreadcrumbList",
"itemListElement": [
{
"@type": "ListItem",
"position": 1,
"name": "Home",
"item": "https://geosnap.ai/en/"
},
{
"@type": "ListItem",
"position": 2,
"name": "Documentation",
"item": "https://geosnap.ai/en/docs"
},
{
"@type": "ListItem",
"position": 3,
"name": "How We Generate Questions",
"item": "https://geosnap.ai/en/doc/how-questions-are-generated"
}
]
}
TechArticle describes the guide
{
"@context": "https://schema.org",
"@type": "TechArticle",
"headline": "How We Generate Questions",
"url": "https://geosnap.ai/en/doc/how-questions-are-generated",
"inLanguage": "en",
"publisher": {
"@id": "https://geosnap.ai/#organization"
},
"isPartOf": {
"@type": "WebPage",
"url": "https://geosnap.ai/en/docs"
},
"description": "The quality of Geosnap's analysis depends on the quality and coverage of the generated questions. The system employs a structured approach that simulates"
}
Metadata
The information in the page header, which crawlers and engines read before the content.