How to Get Found by AI: A Practical Guide (and What You Can Skip)

When you ask Google a long, specific question today, the first thing you often see is a summary written by an AI, with a few links to read more. The same happens with chatbots that search the web. For anyone with a website, a store or a business, the question has changed: how do you show up in those answers?

This guide puts in order what the official documents from Google, OpenAI and Anthropic say, what is actually worth doing and what isn't needed. Some tips circulating online are not backed by the official sources, and we point those out.

What is "GEO" and why does everyone talk about it?

GEO stands for Generative Engine Optimization, meaning optimizing your content for engines that generate answers. It is the name the industry has given to the idea of getting cited by AI.

The most useful news is that, for Google, it is not a separate discipline. The AI features of its search, meaning AI Overviews and AI Mode, use the same eligibility requirements and the same ranking signals as regular search. In practice, good SEO is still the starting point.

What Google says, in short

  • To appear, a page must be indexed and eligible to be shown with a snippet: these are the normal technical requirements of search.
  • No special optimization is needed: according to the official guide, you don't need new files for AI, dedicated text or special markup, and there is no specific schema.org for these features.
  • Meeting all the requirements does not guarantee that you will be crawled, indexed or shown.
  • The AI features can run several related searches to find a broader set of pages. So a page that doesn't answer the exact question, but answers part of the problem, can be cited too.

What to actually do: eight checks

  1. Your site must be reachable. Check that your robots.txt file, your CDN and your hosting don't block the search crawlers.
  2. Write useful content made for people. Original, complete, and adding something to what you can find everywhere else.
  3. Put the important information in the text. Not only in images or videos: those help, but the content must also be there in writing.
  4. Answer specific questions. People now type long, detailed requests. A page about a precise case (for whom, for what use, with what limit) has a better chance than a generic page. This is common-sense advice, not an official rule.
  5. Link your pages together. Internal links make your content easier to find, for people and for search engines.
  6. Take care of the page experience. Fast, readable on a phone, with no obstacles between the visitor and the content.
  7. Keep structured data consistent with the visible text. If you sell online, also keep Merchant Center and your Google business profile up to date.
  8. Make it clear who writes and how. A real author with a short bio is better. And if you use AI to produce content, say so where readers would expect it.

How to control AI crawlers

Different companies use different programs (called crawlers) for different purposes: indexing sites for search, collecting material to train models, or reading a page because a user asked for it. For each one you can decide what to allow in your robots.txt file.

CompanyNameWhat it is for, according to the official documentation
GoogleGooglebotCrawling for Google Search, including the AI features
GoogleGoogle-ExtendedLets you choose whether your content can be used to train future Gemini models. It does not affect your presence or ranking in search
OpenAIOAI-SearchBotShows sites in ChatGPT's search features
OpenAIGPTBotCollects content that may be used to train models
OpenAIChatGPT-UserVisits pages when a user asks it to. According to OpenAI, robots.txt rules may not apply to these visits
AnthropicClaudeBotCollects public content for training
AnthropicClaude-SearchBotIndexes content to improve the search results shown to users
AnthropicClaude-UserVisits sites when a user's question requires it

An example robots.txt that allows search but not training:

User-agent: OAI-SearchBot
Allow: /

User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: Google-Extended
Disallow: /

This is not advice to copy: it is your choice. Blocking the search crawlers can reduce your visibility in those services, while blocking the training ones, according to their respective documentation, does not remove your site from search. Google also says it plans to support a new signal for training preferences in early 2027.

How to measure results

Traffic that comes from Google's AI Overviews and AI Mode is included in the Search Console performance report, under the "Web" tab. According to Google, people who arrive from these results tend to stay longer on the site: that is the company's claim, and you should check it against your own data. That's why, besides visits, it's worth looking in Google Analytics at what people do after they arrive: pages viewed, sign-ups, purchases.

Things you don't need, or that have no confirmation

  • Tricks and text written "to please the AI." For Google's AI features there is no special markup or file.
  • Producing large amounts of AI text with no human oversight. Google treats this as a low-value signal, and using automation to manipulate rankings goes against its anti-spam policies.
  • Changing dates to look up to date. Google lists this among behaviors aimed at search engines, not at people.
  • Expecting guaranteed results. No official source promises that you will appear in the answers.

Frequently asked questions

If I block GPTBot, do I disappear from ChatGPT?

According to OpenAI, the settings are independent: GPTBot covers model training, while OAI-SearchBot covers appearing in ChatGPT search. You can allow one and block the other. Changes for search can take about 24 hours.

Does Google-Extended affect my ranking on Google?

No. Google writes that it does not affect your site's presence in search and is not a ranking signal. It concerns the use of content for Gemini models.

Do I need a special file for AI?

Not for Google's AI features: its official guide says so. For other services, check each one's documentation.

How long does it take to see results?

There is no certain timeframe. For example, Google says that a new crawl, which is needed for preview controls to take effect, can take anywhere from days to months.

Official sources

This article was prepared with the help of artificial intelligence tools, based on the official documentation, and is reviewed before publication.

Condividi questo articolo sui tuoi socialShare this article on your social networksComparte este artículo en tus redes socialesTeile diesen Artikel in deinen sozialen Netzwerken

No products were found matching your selection.

Iscriviti alla newsletter di You vs AISubscribe to the You vs AI newsletterSuscríbete al boletín de You vs AIAbonniere den Newsletter von You vs AI