Visibility in AI search is earned through solid implementation of traditional search engine optimization: the page must be crawlable and indexable, robots.txt must permit the relevant crawlers, and the content must provide an original, verifiable answer to the question asked. Google’s official documentation explicitly states that AI Overviews and AI Mode require no special markup or AI-specific file; the only condition added to Search’s technical requirements is that the website be included under Search Console’s generative AI setting, which is also the default selection. Microsoft likewise states that Bing and Copilot experiences rely on the same crawling, indexing, and ranking foundations as traditional search.
This guide tests recommendations marketed as “GEO” against official sources. It relies on Google Search Central, OpenAI’s crawler documentation, and Bing Webmaster Guidelines; it does not include unsourced lists of “AI ranking factors.” Each section contains an actionable check or decision.
How does AI search select source pages?
Generative AI search features retrieve pages from the search index when constructing an answer and link to pages supporting that answer. Google describes this method as retrieval-augmented generation (RAG, grounding): core ranking systems first find relevant, current pages, and the model then writes the response using information from those pages. On Bing’s side, the same process is called “grounding,” and citations in Copilot answers come from this step.
The second important concept is query fan-out. The model performs several additional related searches that divide the user’s single question into subtopics and retrieves separate results for each. As a result, a page can be cited as a source if it provides the best answer to a subquestion, even when it does not rank highly for the main query. Google also notes that AI Overviews and AI Mode may use different models and techniques, and the links listed in the two features may differ.
GEO and AEO terminology: what do Google and Microsoft say?
GEO (generative engine optimization) and AEO (answer engine optimization) describe work intended to earn citations in AI answers; according to Google, this work is still SEO. Google’s generative AI optimization guidepublished in May 2026 says that optimizing for generative search means optimizing for the search experience and recommends evaluating third-party “AEO/GEO” advice against the official guide.
Microsoft, meanwhile, has incorporated the term GEO directly into its guidelines. Bing defines GEO as focusing on making content eligible for grounding and citation in AI responses; the same passage emphasizes that SEO does not guarantee rankings and GEO does not guarantee citations. The two companies’ approaches meet at the same practical point: there is no need to build a separate “AI website” for visibility in AI search; the technical and content quality of the existing website are decisive. How these principles are implemented within SEO work is explained on the AI search optimization page.
Technical requirements for Google AI Overviews and AI Mode
For a page to be a supporting link in AI Overviews (AI Bakışları in the Turkish interface) or AI Mode, it must be indexed, eligible to appear with a snippet in search results, and included under the website’s “Generative AI in Search” setting in Search Console; inclusion is the default, and no special markup or file is required beyond this. Even when these conditions are met, Google does not guarantee crawling, indexing, or display. The following points should be checked:
- Does the page appear as indexed in Search Console’s URL Inspection tool?
- Does the page have nosnippet or a very short max-snippet value? These values limit the text that can appear in the summary.
- Are robots.txt, the CDN, or hosting-layer bot protections blocking Googlebot?
- Is the main information present as text on the page rather than inside an image?
- Does the structured data match the visible text on the page exactly?
- Does the “Generative AI in Search” control under Settings in Search Console allow the website to be included in these features?
The last item is new. As of August 31, 2026, Google has made the Search generative AI control available to all websites. Inclusion is the default selection; links and content from an excluded website do not appear in AI Overviews, AI Mode, or generative features in Discover, but the website continues to appear in regular search results. This setting manages visibility in search features rather than model training; Google-Extended remains a separate control for training.
The difference between OAI-SearchBot and GPTBot for ChatGPT search
On OpenAI’s side, visibility in AI search begins with allowing OAI-SearchBot; GPTBot is a separate crawler that crawls for model training. OpenAI’s crawler documentationstates that these settings are independent: a website can allow OAI-SearchBot while blocking GPTBot. Websites blocking OAI-SearchBot are not displayed in ChatGPT search answers and may appear only as navigational links. OpenAI also adds that a robots.txt change may take approximately 24 hours to be reflected in search systems.
| User agent | Purpose | What happens if it is blocked? |
|---|---|---|
| OAI-SearchBot | Crawls websites to be displayed in ChatGPT search features | The website does not appear as a source in ChatGPT search answers |
| GPTBot | Crawls content that may be used to train generative AI foundation models | Communicates a preference against using content for model training; does not independently determine search visibility |
| ChatGPT-User | Visits the page when a user asks a question or initiates an action | Because it is user-initiated, robots.txt rules may not apply; it does not determine appearance in search |
| Googlebot | Supplies Google Search and the AI features within search | The page cannot be crawled; its content cannot be used in either traditional results or AI Overviews |
The last row of the table prevents a common misunderstanding. Google has no separate search crawler for AI features; blocking Googlebot means the page’s content cannot be used in any part of Google Search.
Page contents
Checking robots.txt: are crawlers being blocked accidentally?
For websites that never appear in AI search engines, the first places to examine are the robots.txt file and server-side bot filters. Security plugins, “block AI bots” options in CDN services, or an old robots.txt template can create restrictions the website owner is unaware of. The check proceeds in this order:
- Open alanadi.com/robots.txt in the browser and read the blocks mentioning Googlebot, Bingbot, and OAI-SearchBot.
- Verify that general Disallow rules under “User-agent: *” do not block important folders.
- Review automatic bot-blocking settings in the CDN and firewall dashboards; this layer can reject requests even when robots.txt permits them.
- Check server access logs to see whether the relevant user agents receive 200 or 403 responses.
- Run a live test of an important page in Search Console’s URL Inspection tool.
An example configuration for a website that wants to preserve search visibility while limiting model training is shown below. This is only an example; which crawlers to allow is determined by the business’s content policy.
User-agent: OAI-SearchBot
Allow: /
User-agent: GPTBot
Disallow: /A website-wide audit of crawling and indexing settings is handled within technical SEO services .
Let’s check crawler access together
When you send the website address through WhatsApp, let’s review robots.txt, bot protection settings, and indexing status together in a free preliminary assessment. No promise is made that the website will be cited in AI answers; the restriction that needs to be removed becomes clear.
Bing and Copilot: what do Microsoft’s guidelines require?
To be cited in Copilot, a website must be represented accurately and currently in Bing’s index; Bing Webmaster Guidelines list the ways to achieve this item by item. The guidelines require content changes to be reported promptly to reduce outdated or incorrect addresses in Copilot answers. The effects of robots meta directives on AI answers are also spelled out this clearly for the first time:
- IndexNow: Newly added, updated, or deleted addresses are reported to Bing immediately; reporting as changes occur is recommended instead of submitting in batches.
- Sitemap: Contains only canonical addresses, removes deleted or redirected addresses, and uses a lastmod value reflecting the actual update date.
- NOARCHIVE: Prevents content from being used in Copilot answers and grounding results.
- NOCACHE: Allows Copilot to use only the address, title, and snippet from the page; the depth of citation is reduced.
- NOSNIPPET and data-nosnippet: Prevent Bing from showing description text and may limit the quality of Copilot citations.
- Prompt injection: Content intended to steer or interfere with language models used by Bing and Copilot, whether hidden or visible, can lead to lost visibility or removal from the index.
Another notable principle in Bing’s guidelines is to place important information near the top of the page and focus each address on a single main topic. Pages that state the result in the first paragraph are considered more suitable for grounding than pages that reach the answer only after long introductions; for visibility in AI search, this is the least expensive improvement to page structure.
Is an llms.txt file necessary?
It is not necessary for Google Search; Google’s official guide explicitly states that it does not use AI-specific text files, including llms.txt, special markup, or Markdown versions. According to the same document, creating this file neither harms nor helps visibility on Google because Google Search ignores these files. There is no problem creating it if another service or system uses it.
OpenAI and Bing documentation likewise defines no support for llms.txt; both sources explain access management through robots.txt and robots meta tags. When visibility in AI search is the objective, fixing robots.txt, the sitemap, and page content before spending effort on llms.txt provides a more concrete return.
Common characteristics of cited content
Content cited in AI answers offers information, experience, or a perspective unavailable elsewhere and can be understood on its own. Google describes this as “non-commodity content”: content based on firsthand experience and expertise rather than articles repackaging information everyone already knows. Bing requires addresses preferred for grounding to be independently verifiable. The practical implications of these two principles on a page can be summarized as follows:
- Definitions and basic facts are stated clearly; an important sentence can be understood without referring to the previous paragraph.
- Business, product, service, and location names are used consistently throughout the website.
- Each section answers the question in its heading in the first sentence.
- Numbers, dates, and technical claims are provided with their sources.
- Images and videos support text rather than replace it; alternative text and captions are added.
- When information changes, the page is updated and the update is reported to search engines.
In the same guide, Google warns that producing a separate page for every variation a searcher might ask can constitute scaled content abuse. A smaller number of in-depth pages also provides a stronger foundation for AI search engines. How to build a content plan using this approach is explained on the content marketing page.
Additional steps for local businesses and e-commerce
For businesses providing local services, visibility in AI search depends on completing the Business Profile accurately and fully alongside the website. Google states that generative answers can include local business information and product listings where appropriate and recommends keeping Business Profile and Merchant Center data current for this purpose. Opening hours, service area, category, and contact information must match the website.
For e-commerce websites, ensuring that prices, availability, and descriptions in the product feed match product pages continues the same principle. The steps for editing a profile are detailed in the Google Business Profile optimization guide .
How do you measure visibility in AI search?
Official measurement tools now exist: Google Search Console’s generative AI performance report and the AI Performance dashboard in Bing Webmaster Tools. Google’s report collects impressions in AI Overviews and AI Mode separately from traditional results; it provides impression counts only, with no click or query data. Data can be examined by page, country, device, and date, and the report has been available to all websites since August 31, 2026. Bing’s dashboard, launched in preview in February 2026, lists how often content is cited in Copilot and Bing AI summaries, which pages are cited, and a sample of grounding queries that produced citations.
| Measurement tool | What does it show? | Limitation |
|---|---|---|
| Search Console, generative AI performance report | Impressions in AI Overviews and AI Mode, by page | Counts impressions only, without clicks or queries; the report may not appear if impressions are insufficient or the website is excluded from these features |
| Search Console, performance report (Web) | All search traffic, clicks, and queries, including AI features | Collected together with traditional results |
| Bing Webmaster Tools, AI Performance | Citations and cited pages in Copilot and Bing AI summaries | Query data is sampled; it provides no ranking or position information |
| Manual query checks | Whether the brand and page are mentioned for specific questions | Answers vary by person, time, and location; a one-time observation cannot be generalized |
Google also warns against third-party tools claiming access to internal ranking or AI metrics; no external tool has access to these systems. Microsoft reminds website owners that declining clicks do not always mean lost visibility, since content can appear as a citation without receiving clicks.
Common mistakes and the right approach
The most common mistake is spending time on “tricks” unsupported by official documentation while postponing fundamental deficiencies. Google’s list of myths says there is no need to split content into tiny chunks for AI, adopt a writing style exclusively for AI, or artificially collect “mentions.” Structured data is not mandatory for these features either; it remains useful for rich results, but adding a special schema.org type does not deliver visibility.
The second group of mistakes is technical. Leading examples include bot protection blocking search crawlers too, important information existing only in an image or a component loaded later on the client side, and adding nosnippet to pages and then forgetting it. The correct sequence for visibility in AI search is first to resolve access and indexing problems, then make the pages receiving the most questions clearer and better sourced, and monitor results through official reports. The overall framework of SEO work is on the SEO service page, and the basic checklist appears in the SEO-friendly website guide .
Let’s set priorities for AI visibility together
For businesses just researching the subject, weighing priorities, or ready to implement changes immediately, the starting point is the same: seeing the current state of access, indexing, and content clarity. Share your website address so we can decide together in a free preliminary assessment which issue to address first; a clear quote is shared once the work items are defined.
Related guides
Frequently asked questions
Is separate optimization required for visibility in AI search?
According to Google, it is not; there are no additional technical requirements or special markup for AI Overviews and AI Mode. The fundamental conditions are that the page be indexed, eligible to appear with a snippet, and contain original content.
What is the difference between GEO and SEO?
GEO is the name for work focused on making content eligible to be cited in AI answers; Google considers this work part of SEO. Bing also defines GEO but states that it does not guarantee citations.
Does adding an llms.txt file increase visibility on Google?
No; Google states that Search does not use llms.txt files and that these files neither help nor harm visibility. The file can be created if another service uses it.
Does blocking GPTBot prevent a website from appearing in ChatGPT search?
This does not cause that outcome on its own; OAI-SearchBot is used for ChatGPT search results, and the two settings are independent. If OAI-SearchBot is blocked, the website is not cited as a source in ChatGPT search answers.
Where can impressions from AI Overviews be seen?
Impressions in AI Overviews and AI Mode are tracked separately in Search Console’s generative AI performance report. This traffic is also included in the Web search type of the general performance report.
Does adding structured data make a website appear in AI answers?
It does not; Google states that no special schema.org markup is required for generative AI features. Structured data remains useful for rich results and must match the text visible on the page.
When will changes be reflected in AI answers?
There is no fixed timeframe; according to Google, recrawling a page can take anywhere from a few days to a few months. OpenAI states that robots.txt changes may take approximately 24 hours to be reflected in the search system.
Call us
Call us for detailed information.
Let us call you!
Share your phone number and let us call you!
Our other services
- › Corporate web design
- › E-commerce website development
- › SEO optimization
- › WordPress website development
- › Google Ads management
- › WordPress speed optimization
Service areas
- › Web design in Sakarya
- › Web design in Tekirdağ
- › Web design in Bursa
- › Web design in Beşiktaş
- › Web design in Şişli
- › Web design in Tuzla
- › Web design in Ümraniye
- › Web design in Kadıköy
- › Web design in İstanbul
- › Web design in Ankara
- › Web design in Ataşehir
- › Web design in Üsküdar
- › Web design in Kartal
- › Web design in Maltepe
- › Web design in Pendik
- › Web design in Kocaeli
- › Web design in İzmir
- › Web design in Antalya
