What GEO is and why a business needs it

Each assistant has its own exact order, and it is not published. Assistants with search follow roughly the same scheme.

• It searches. From the user's question the assistant makes one or more search queries and reads the pages it finds.
• It picks passages. The answer takes the pieces that answer the question directly: a definition, a list of steps, a table of prices and timelines.
• It combines sources. Besides your site the assistant reads directories, maps, industry articles and reviews. If the name, services and contacts differ between them, the answer may carry a mistake.
• It skips what it cannot read. Many assistants' crawlers do not run scripts, so text that appears only after scripts load is empty to them. Prices shown only in images and pages closed to these crawlers in robots.txt are lost the same way.

• Access for crawlers. robots.txt lets the assistants' crawlers in, pages are served as ready HTML, and there is a sitemap. GPTBot collects data to train OpenAI models, while ChatGPT search uses OAI-SearchBot, so you can close the first and keep the second.
• Answers to clients' questions. Service pages and guides open with the answer in two or three sentences, prices and timelines are written as text, and there are tables and a questions section. The guides on this site are written that way too.
• schema.org markup. It repeats in machine terms what the page says: who you are, your services and starting prices, your articles and answers to questions.
• An llms.txt file. It is a short plain-text map of the site for AI models. The format is not a standard yet and not every assistant reads it, but it takes little time to make.
• Mentions on other sites. Profiles in directories and on maps, the same name, address and contacts everywhere, articles on industry sites, client reviews after finished work.
• Checking. Once a month the assistants are asked the questions clients ask, and the answers are logged: whom they name and which pages they link to.

• Ask ChatGPT, Perplexity and Gemini five to ten questions your clients ask, with your city: “where to order a website in Almaty”, “what a Telegram bot costs”. Write down whom they name and which sites they list as sources.
• Ask about the company itself: what it does and what its services cost. Mistakes in the answer show which facts about you are missing from public sources.
• Open yoursite/robots.txt and check that OAI-SearchBot and PerplexityBot are not blocked, and that a “Disallow: /” line under “User-agent: *” does not block every crawler at once.
• Turn off JavaScript in your browser and open the site. If the text and prices disappear, crawlers that do not run scripts will not see them either.
• In Metrica, look for visits from chatgpt.com, perplexity.ai and gemini.google.com in the traffic sources report. Whether they led to payments is shown by end-to-end analytics.

• A business whose clients compare and ask questions before they buy: services, B2B, expensive goods.
• Companies whose site is already indexed and gets impressions in search. GEO uses the same base and adds work on answers and mentions.
• It is too early if the site is not indexed by Google and Yandex, or opens empty without scripts. The technical base comes first, otherwise the assistants have nothing to read.

Fill in the brief: four short steps, about three minutes. Give the site's address and the questions clients should find you by. Within a working day we send a work plan and an estimate with the price and the timeline.