iQ Centre Case Study: AI Search Visibility Growth from 16% to 39%
GEO · AI SEO · Evolution CMS · Schema.org
Prepared the site of the educational network iQ-center for more unambiguous reading by search robots and AI-systems: organized technical URL, linked entities in JSON-LD and described the regional structure in llms.txt. After the completion of Share of Voice in Alice, AI grew from approximately 16% to 34X11QXZZX12QXZ%.
Project and task
iQ Center — is a federal network for preparing schoolchildren for the exam and OGE. The site has regional centers, online training, subject areas, teachers, graduate results and organization documents.
It is not enough to open pages for indexing. It is important that the system correctly links the service with the city, distinguishes online learning from face-to-face classes and finds the main source of each fact.
What needed improvement
- Separate useful pages from CMS service sections.
- Reduce the ambiguity between similar URL.
- Describe the organization, pages and supporting documents.
- Clearly distinguish between cities and learning formats.
How to evaluate the result
Compared weekly values of Share of Voice in Alice AI before and after completion of work. Source of observations — statistics Yandex Webmaster, given in the source materials of the case.
Work completed August 1, 2026. Observation period: 29 June to 24 August 2026
What did I do?
1. Cleared the circumvention rules.
In robots.txt excluded service directories Evolution CMS: backups, cache, export and import, modules, plugins, snippets and installer. Required to display pages CSS, JavaScript and images left available.
Added an explicit link to the site map: sitemap.xml.
2. Ordered sitemap and canonical
I put in order the site map and canonical addresses. The task of — is to identify the main versions of pages and reduce the competition of similar URL in the structure with cities, exams and subjects.
So each page has a more understandable place in the overall architecture of the site.
3. Connected entities through Schema.org
Prepared JSON-LD with related @graph. The organization is described as EducationalOrganization— website as WebSite, the selection page of — as CollectionPage.
Own @id connections about, isPartOf allow you to clearly indicate which organization you are talking about and to which site the page belongs. Navigation describes BreadcrumbList.
4. Described the educational license
Linked the organization through hasCredential singularly EducationalOccupationalCredential. The license has an identifier, name, category and address.
PDF with an extract from the registry is presented as DigitalDocument. The result is a clear chain: the organization → license → confirming document.
5. Prepared by llms.txt
Gathered. llms.txt as a compact guide to the site: areas of training, subjects, online learning, teachers, results, graduates and branches.
He added the rules of interpretation: check current prices and schedules on the profile page, do not transfer addresses between cities and do not interpret graduate results as a guarantee.
6. Separate regions and online
He described the model of the regional section: the city of → USE or OGE → subjects, teachers and results. For Moscow designated the Moscow section, for remote format — page /online, /online-ege and /online-oge.
The goal of — is to provide additional context so that information about one branch is not mixed with information about another.
Result: Share of Voice increased to 34XQ4QXZ39%
Five weeks before the completion of work, the indicator remained approximately at the level of 16%. At the first weekly point after completion, August 3, it was about 35%. For the next three weeks, values remained in the range of 34–XQ4QXZ%.
| Date. | Share of Voice | Phase |
|---|---|---|
| 29 June | ≈16% | Until completion of the work |
| July 6 | ≈16% | Until completion of the work |
| 13 July | ≈16% | Until completion of the work |
| 20 July | ≈16% | Until completion of the work |
| 27 July | ≈16% | Until completion of the work |
| August 3 | ≈35% | After completion of the work |
| 10 August | ≈34% | After completion of the work |
| 17 August | ≈39% | After completion of the work |
| 24 August | ≈36% | After completion of the work |
At the peak, the difference was about 23 percentage points: 39% vs 16%. In the four weeks after the work, the figure was about 2.1–2.4 times higher than the baseline.
Webmaster status
Your site is in the top 3 by mentioning » — for queries where the site ranks highly in organic Search. This status refers to an observable set of queries, not all of Alice’s possible responses.
What this case confirms
After comprehensive technical and semantic optimization, the measurable visibility of the site in Alice AI has increased. There was no separate experiment for robots.txt, Schema.org, or llms.txt: the changes were implemented together.
So I don’t attribute growth to a single file, and I don’t consider time matching as proof of each tool’s contribution. This case also did not measure changes in visibility in ChatGPT, number of applications or revenue.
Next step — answers real questions
The described set of works did not include a separate processing of FAQ on the site iQ-center. The next step of — is to collect user information requests and prepare answers according to the scheme: a specific question → short answer → explanation → confirming the source.
After such an update, it is possible to continue observations and compare the dynamics with the previous period.
Frequent questions
What was done for iq-centr.ru visibility in a AI search?
I organized robots.txt, sitemap and canonical, and prepared the associated Schema.org and llms.txt markup with region descriptions and interpretation rules. The work affected the technical structure and presentation of facts on the site.
What is the result of Alice AI?
After completion on August 1, the 2026 Share of Voice grew from approximately 16% to 34XQ6QXZZ39%. In four weekly observations from August 3 to 24, the indicator remained above baseline. The webmaster showed the status of the top 3 mentions for queries, where the site ranks high in organic Search.
Can growth be attributed to llms.txt alone?
Nope. llms.txt was added along with other changes, so its individual contribution was not measured. The result refers to the period after the work, not to a single file.
Why describe the license in JSON-LD?
In this project, markup links the educational institution, its license and PDF with an extract from the register. So the supporting document is included in the explicit structure of facts about the organization.
Why are separate rules important for a website?
Addresses, prices and schedules can vary between cities. In llms.txt, I identified regional sources and separately described online destinations to provide context for the correct use of information.
Was FAQ recycled at the iQ site?
No, a separate reprocessing of the FAQ was not part of the described step. It's planned as the next step. The questions on this page explain the case itself and are not the result of the customer’s FAQ update.
Can you guarantee the same growth for another website?
Nope. Here are observations on one project and one period. For another site, you need to evaluate the original structure, implement appropriate changes and measure the dynamics separately.
Need technical preparation of the site for AI search?
I will help you understand the structure of URL, indexing, entities and regional sections. You can start by auditing the site and selecting indicators to evaluate the result.