Sample Report · Individual Measurement · Content · September 2026
You are mentioned. Cited nowhere.
Service provider with a WordPress site of over 800 pages, active in nine English-speaking countries. Content measured using a site-wide crawl and a deep measurement on the EQwise SEO platform. Additionally, a usability panel based on eight of the client’s own buyer profiles, and AI responses from ChatGPT with web search across nine countries. Name, domain, page names, and field of expertise have been omitted; every number is an actual measurement.
largest source per country50
the service provider itself0
Average across nine countries for the direct purchase inquiry, own calculation from the measurement on September 21, 2026.
The core in two numbers
AI responses
0 out of 324
with a clickable reference to the service provider, nine countries
Buyer profiles
4 out of 8
drop off due to a missing or unclear price
The site has sufficient contentContent All content on your website, such as text, images, and videos.: over 800 pages, actively maintained and updated. The problem is not the amount of text.
It lies in three places where that text does not arrive. First, in the structure that a search engineSearch engine A website used to search the internet, such as Google or Bing. or an AI systemAI Assistant A chat program such as ChatGPT, Claude, Gemini, or Perplexity that you can ask questions in natural language. must be able to read. Second, with the buyer who looks for a price and finds nothing. Third, with the AI assistant that looks for a source to cite and chooses a page other than the service provider’s own.
Three places where the text does not arrive
The same measurement viewed from three perspectives. What an AIAI Computer programs that perform tasks normally requiring human intelligence, such as writing text or answering questions. response engineEngine The system that generates answers or search results, such as Google or ChatGPT. does and does not cite, what a buyer misses on the page, and how the site is structured.
AI responses
The service provider is well-positioned with search engines. In AI responsesAI Answer The generated response provided by an AI assistant in answer to a query. to purchase inquiries, the picture is different. Across 324 responses in nine countries, measured on September 21, 2026, the site never received a clickable reference.
In an earlier, smaller measurement of 120 responses in five of those countries, the external overview page was cited 445 times as a source. For the service provider itself, that was zero. The service provider is mentioned occasionally; cited with a link, never.
The ratio in the meter chart at the top (50% versus 0%) comes from the same measurement on September 21, 2026. On average across nine countries, the largest source per country—usually the external overview page—accounts for 50% of the citations for the direct purchase inquiry. This is an own calculation based on the nine individual country measurements. The service provider itself stands at 0%.
The remaining 50% is spread across 4 to 13 domains per country. For a broader, more open question about the same field, the share of the largest source drops to an average of 11% (own calculation). The citations are spread across 18 to 28 domains per country. For that question, there is not yet a dominant source.
A likely explanation lies in the data modelData Model The way data is organized and interconnected within a system., not in the text. Price and duration are recorded on the service pages as a category, not as a field, meaning a valid “OfferOffer A component of structured data in which an offer including price is recorded.” is missing from the page’s structured dataStructured data Additional information in the code that tells search engines and AI exactly what is on a page.. That is exactly the field an AI response engine can read to cite a service with a price and date.
That this plays a role is supported by what the assistant does do. When asked for a provider of this service, it mentionsMention A place where your business name is mentioned, with or without a link. multiple parties and links them all to the same external overview page.
That page is explanatory and well-organized, with a fixed spot for every party mentioned. The provider’s own service page lacks those fixed fields, suggesting that is why it is skipped. This is an explanation, not a separate measurement of the cause itself.
The service provider is also listed on that page, just mentioned slightly differently than how the service is currently delivered.
Why this is not a story of despair. This suspected mechanism can be managed: a valid price field on your own page and a correct listing on the external overview page do not require a rebuild. For the broader question without a dominant source, the situation is reversed. No one is the fixed reference yet, which is as much an opportunity as a risk.
Search intent
Fifteen service pages promise answers to a buyer’s most important choice-related questions in their headings but leave that promise unfulfilled. A usability panel based on eight of the client’s own buyer profiles shows the cost of this. Four of the eight profiles drop off due to exactly a missing or unclear price, part of the information missing from those fifteen pages.
| Buyer | Drops off at | Strength of evidence |
|---|---|---|
| 1 | a price without a payment plan | strong |
| 2 | a price that is nowhere on the page | strong |
| 3 | “price on request” instead of an amount | strong |
| 4 | unclear conditions to start | moderate |
| 5 | a price that must be requested separately | moderate |
| 6 | a promise without information about quality | moderate |
| 7 | jargon | weak |
| 8 | a tone that assumes more prior knowledge than this buyer has | weak |
Structure
The site consists of five types of pages. The distribution itself is healthy: maintained and complete. The problem is not in the numbers but in two areas beneath them. It concerns how a service is recorded, and whether a page has a subject that a search engine or an AI system can read.
| Page type | Number | Judgment |
|---|---|---|
| Pages | 201 | leave as is maintained, robots and canonicals in order |
| Blog posts | 429 | review only counted, content not yet reviewed for substance |
| Knowledge articles | 39 | leave as is no separate finding |
| Sellable services | 15 | replace separate from the descriptive page; at least one service exists two to three times in the system |
| Taxonomy archives | 126 | leave as is correctly listed in the sitemap |
The second layer is the heading structure, and it is less healthy than the numbers. Of five core pages checked separately, not a single one has a real main title (H1) (re-tested on September 20, 2026: still zero on all five). A broader platform audit counts this site-wide for 54 of the 834 crawled pages.
A visitor won’t notice this: the heading appears large and clear on the screen. For a search engine or an AI system, such a page has no defined subject.
| Finding | Number | Judgment |
|---|---|---|
| No H1, site-wide count across 834 pages | 54 | fix one line in the template |
| No H1 or more than one, deep measurement on 20 pages × 2 widths | 32 | fix same cause, inexpensive to restore |
| Five core pages checked separately for H1 | 0 out of 5 | fix none of them have one |
Opportunities
Five interventions follow directly from the measurement, from small to large.
| Intervention | Judgment | Scope |
|---|---|---|
| Add price and time indication to the fifteen service pages | repair | small, repeated adjustment per page |
| Restore heading structure (H1 per page) | repair | one line in the template |
| Correct listing on the external overview page | repair | outside own site |
| Price and duration from category to field, with a valid Offer in structured data | replace | part of a foundational step; affects the data model |
| Clean up duplicate recording of the same service | replace | part of the same foundational step |
The first two can be done immediately and directly affect the purchase decision and visibility. The last two affect the same underlying data model and therefore belong together. As long as price and duration are categories instead of fields, every page remains invisible to the field an AI response engine reads. This applies regardless of how much text is on it.
What meets the standard
Here too: what is correct, is left alone.
- The sitemap is complete and up-to-date: seven pages modified in September, nine in August.
- Robots and canonicals are correct on all five measured core pages, and no AI crawlers are blocked.
- Five previously reported broken service links were restored by the client themselves with redirects, without being asked.
- The editorial team is active: it is the structure beneath the pages that is stagnant, not the maintenance of the content.
What this does and does not prove
What it proves
- The inventory, heading structure, and the lack of an H1 on the five core pages were measured and can be reviewed on the dates mentioned.
- AI responses to purchase inquiries do not link the site, while the external overview page is cited. This was measured across 324 responses (September 21, 2026) and an earlier set of 120 responses.
- That four of the eight buyer profiles drop off due to a missing or unclear price was measured with a usability panel on the client’s own buyer profiles.
What it does not prove
- The quality of the 429 blog posts; these were only counted, not qualitatively assessed.
- That a correction on that overview page alone will solve AI visibility; the link profile and data model also play a role.
- That every visitor experiences this exactly the same way; that depends on the question someone asks and the country in which they search.
Measurement setup and sources
Site-wide crawl of 834 pages (platform audit), EQwise SEO platform. Deep measurement on 20 pages × 2 widths. Usability panel on eight of the client’s own buyer profiles.
AI visibility measured with ChatGPT with web search enabled, country as search location. This involves 324 responses across nine countries, measured on September 21, 2026. Additionally, there is an earlier, smaller set of 120 responses in five of those countries.
This is an anonymized sample of an actual measurement for a client; name, domain, page names, and field of expertise have been omitted or replaced by a description.
How many AI responses cite your site?
A content measurement and an AI visibility check (SEO and GEO) reveal where the text misses a buyer. They also show where an AI response engine cites a source other than yours.
Request the SEO and GEO auditMore about the SEO and GEO audit
