--- title: "GeoScan: case study" description: "GeoScan: in 10 seconds shows whether your site is ready to be cited by AI and whether AI knows your brand." url: https://seo7.es/en/proyecto/geoscan canonical: https://seo7.es/en/proyecto/geoscan lang: en type: CreativeWork --- # GeoScan: case study > GeoScan: in 10 seconds shows whether your site is ready to be cited by AI and whether AI knows your brand. Page: https://seo7.es/en/proyecto/geoscan ## Summary for AI GeoScan: in 10 seconds shows whether your site is ready to be cited by AI and whether AI knows your brand. People stopped googling and started asking. A question like "recommend a groomer nearby" or "good car repair shop in Valencia" gets a ready answer with company names, and in that answer your business is either there or not. ## Questions and answers ### What does GeoScan by SEO7 check? GeoScan is SEO7's free tool at geoscan.seo7.es. It shows whether ChatGPT, Perplexity, Gemini, Claude and Copilot can find, read and cite a page of your site, or will name a competitor instead. The check covers 172 points in 8 groups, and the result comes as two independent numbers, with no sign-up or installation. ### Why does GeoScan give two numbers instead of a single score? The first number shows whether a page is ready to be cited. The second shows whether the model knows the brand before it opens the page. Studies show brand mentions affect citability three times more than links, and the model takes 85% of mentions from third-party sites. A single overall score would hide brand recognition behind technical readiness. ### What lesson did the first version of the GeoScan engine teach? Version 1.1 ran in July and August 2026 and produced a single GEO Readiness Score from a registry of 305 signals in 7 categories. The seo7.es site scored 96 out of 100 and was not cited in any of the six queries. This was the reason for moving to version 1.2. ### How does GeoScan calculate a page's citability? In the citability calculation, page content is multiplied by AI bot access; addition is not used here. If GPTBot, ClaudeBot, PerplexityBot, or Google-Extended cannot reach the page, even good text won't produce a high number. The score runs from 0 to 100 with a grade from A to F, across groups weighted by influence. ### What does an empty brand recognition value in GeoScan mean? Brand recognition is calculated from external signals: mentions, reviews on third-party sites, Wikipedia, Reddit. If there are fewer than three independent external signals, GeoScan returns an empty value instead of a low number. This way the tool honestly shows that there isn't enough data for a conclusion and doesn't pass off a random estimate as an accurate one. ### What is the Reality axis in GeoScan? The Reality axis sends 6 real queries to a language model about the site's niche and checks whether the model named the brand or its competitors. The result shows the specific domains that outrank the site. For seo7.es, for example, cyberclick.es, humanlevel.com and seoestudio.es came back. The queries are handled by the DeepSeek language model. ### Why can an honest 100 points be reached in GeoScan? Inapplicable checks drop out of the denominator and don't penalize the site. A recipe schema on a lawyer's site or a transcript on a page without video don't count toward the score. GeoScan's first scoring penalized everything indiscriminately and 100 points were unreachable, so the logic was rewritten. Now a site is scored only on what applies to it. ### How does GeoScan's full mode differ from the quick analysis? The quick check takes about 10 seconds and examines the page itself: bot access, content, markup, headings. The full mode takes up to two minutes and adds load speed, external brand mentions and a live query to four models. Data from the “Off-site” group, including Search Console, is available only in full mode. ### What does the GeoScan PDF report include? The report contains the score by group, all 172 checks with result and description, the Reality axis and priority fixes ready for a developer. The result appears on screen immediately and for free, and the full PDF report arrives by email in Spanish, Russian or English. A separate Node and pdfmake service renders it. ### Which 8 groups of checks does GeoScan use? Seven GeoScan groups relate to the page itself: access, markup, meta, content, citability, technical and trust. The eighth group, "Off-site", evaluates the brand's presence beyond the site: mentions, reviews on third-party platforms, Wikipedia, Reddit. The base run has 169 checks, and for niches with rich markup the number reaches 172. ### Where does a request left in GeoScan go? The user leaves an email to get the PDF, and the lead with the site address and check results lands in the "Leads" section of the SEO7 admin with the source GEOSCAN. A notification arrives in Telegram and by email at the same time. GDPR-guard checks the consent and the address, and a failure to save the lead doesn't block the email to the user. ### What technical difficulties with n8n had to be solved in GeoScan? n8n Code nodes have no global URL object, so requests silently returned nothing and addresses had to be parsed with regular expressions at every step. Also, n8n unpacks the response and strips the content-encoding header, which gave gzip sites a false penalty. We widened the compression check, and now Vary: Accept-Encoding is accepted too. ### Why is the GeoScan scan split into separate pipeline steps? The pipeline in self-hosted n8n is split into visible steps: loading, parsing, access, citability, markup, content, technical, external signals and Reality. Each step can be opened and run separately, so an error is visible at its own stage together with the intermediate result. There is no Redis on the server, and the scan is built into a synchronous workflow. ### Where is the GeoScan methodology published? The method is described in the «Methodology» section on geoscan.seo7.es in three languages, in a working paper on Academia and in the methodology repository on GitHub under the CC BY 4.0 license. The repository contains a description of the method without source code, weights or prompts. These publications also boost the studio's own visibility, since language models read them. ## Other languages - RU: https://seo7.es/ru/proyecto/geoscan - ES: https://seo7.es/es/proyecto/geoscan ## schema.org markup ```json { "@context": "https://schema.org", "@type": "CreativeWork", "name": "GeoScan: case study", "description": "GeoScan: in 10 seconds shows whether your site is ready to be cited by AI and whether AI knows your brand.", "url": "https://seo7.es/en/proyecto/geoscan", "mainEntityOfPage": "https://seo7.es/en/proyecto/geoscan", "inLanguage": "en-GB", "isPartOf": { "@type": "WebSite", "name": "SEO7.ES", "url": "https://seo7.es/" }, "publisher": { "@type": "Organization", "name": "SEO7.ES", "url": "https://seo7.es/" } } ```