Issue No. 02 · GEO service for Kazakhstanp. 03 · methodAlmaty - Astana - all of Kazakhstan
How we measure
A measurement is not a screenshot of one answer. It is 60‑100 questions, asked the way buyers ask them, under the same conditions and repeated with the same method.
01method
Buyer questions
We collect 60‑100 questions for your niche and city: where to buy, who to turn to, what to choose, how much it costs. We phrase them the way people write, not the way ads are written.
A fresh chat for every question
We ask each question in a new chat, with no history and no personal data. That way the answer does not adapt to past conversations and is closer to what a new buyer will see.
Three platforms
We ask ChatGPT, Perplexity and Google AI. They draw on different sources: ChatGPT search, for example, relies on the Bing index, which is why the checklist includes submitting the website to Bing Webmaster.
We don’t sell a place in the answer. We sell a way to measure it honestly.
What we record
For each question we note whom AI named, in what order and whether you are among them. Together this gives a picture of the niche: how often AI names you and your competitors, and in which questions you are missing.
Website checklist
The audit follows a single list, and you can check every item yourself:
- Prices and availability in the page text, not in an image or a script
- Schema.org markup: products and offers, organisation, FAQs, reviews - if there are any
- llms.txt - a file that tells AI what the business is and where the key information is
- A clean sitemap and robots.txt open to GPTBot, PerplexityBot and ClaudeBot
- Answer pages for the questions people ask AI
- The same company details on the website, Google Maps and 2GIS
- Submitting the website to Bing Webmaster
The same method, repeated
We ask the same questions in the same way 3‑4 weeks after the fixes. Indexes take time to update, so checking earlier would be premature. In the report the two measurements sit side by side: before and after.
What the method does not promise
AI answers are non-deterministic: the same question can get different answers. That is why we do not promise a place or sales growth. We promise measurement using the same method before and after, work to a checklist and a transparent report.
02glossary
Glossary
- GEO
- Generative Engine Optimization - the work of helping AI assistants read and cite a website.
- llms.txt
- a text file at the root of a website: what the business is, which pages matter most, where the prices are. Like a sitemap, but for language models.
- Schema.org
- a shared markup vocabulary that helps search engines and AI tell where the price, address and opening hours are on a page.
03questions
Questions about the method
What is GEO?
Generative Engine Optimization: the work of helping AI assistants read your website and name you when a buyer asks a question.
How is GEO different from SEO?
SEO works with a list of links: a person sees ten websites and chooses for themselves. GEO works with a ready-made answer in which AI has already picked a few names. Both rest on the same foundation - a website that search engines can read.
Why a fresh chat for each question?
So that the answer does not depend on the chat history. Otherwise AI adapts to earlier questions, and the before and after measurements cannot be compared fairly.
Why repeat after 3‑4 weeks?
Search engine and AI indexes do not update instantly. Before then, changes to the website may not have reached the answers yet.
Do you need access to my website for the measurement?
No. We base the measurement and the audit on your public pages - the way AI sees them.
What is llms.txt?
A text file at the root of a website that tells AI what the business is, which pages matter most and where the prices are. Like a sitemap, but for language models.
04get started
Take a look at your niche
Send us «ЗАМЕР»: we will start with the first page of the audit, free.