PLEN
Get in touch

GEO in 2026: how to get ChatGPT and Google AI to recommend your business

More and more customers no longer type keywords into a search engine. They ask an assistant instead: which physiotherapist in Mokotów, which apartment in central Warsaw has parking. ChatGPT, Gemini, Copilot and Perplexity answer with a few names and links, and any business that is not on that short list simply does not exist in the answer. I went through the vendors' own documentation to see where these tools get their information, what they tell site owners to do, and how you can measure your visibility for free as of November 2026. No tricks, because the vendors themselves say there aren't any.

Where AI assistants get information about businesses

When an assistant answers a question about specific services or places, it pulls from a current index of web pages, not only from what the model learned earlier. Google says this plainly: AI Overviews and AI Mode use grounding, meaning they retrieve up-to-date pages from the search index using core ranking systems and show links to pages that support the answer. On top of that, the model issues several related searches on its own (Google calls this query fan-out) to gather more results.

The starting point: if your site is not in the index, or it blocks the assistants' crawlers, no amount of content work will help.

  • Google AI Overviews and AI Mode: the Google Search index. A page has to be indexed and eligible to appear in results with a snippet.
  • Gemini: according to Google's documentation, the Gemini app uses content from the Google Search index for grounding, and site owners control this with the Google-Extended token in robots.txt.
  • ChatGPT: ChatGPT search shows sites collected by the OAI-SearchBot crawler. Sites that block it will not be shown in ChatGPT search answers, though according to OpenAI they can still appear as navigational links.
  • Perplexity: the PerplexityBot crawler is used to surface and link websites in Perplexity's results.
  • Copilot: Microsoft reports citations in Copilot and in Bing's AI summaries inside Bing Webmaster Tools, so it makes sense to make sure your site is present in Bing.

Robots.txt: which crawlers not to block

The robots.txt file tells crawlers which parts of a site they may visit. It happens that a plugin or a hosting setting blocks AI crawlers wholesale. It helps to separate the crawlers that decide whether you appear in answers from the ones that collect content for model training.

  • Googlebot: controls whether you appear in Google Search, including AI Overviews and AI Mode.
  • Google-Extended: controls whether your content may be used to train Gemini models and for grounding in the Gemini app. Google says it does not affect inclusion in Google Search or ranking.
  • OAI-SearchBot (OpenAI): ChatGPT search. OpenAI recommends allowing it. A robots.txt change can take about 24 hours to take effect.
  • GPTBot (OpenAI): collects content that may be used to train models. OpenAI says this setting is independent of OAI-SearchBot.
  • ChatGPT-User (OpenAI): visits a page when a user asks for it. According to OpenAI, robots.txt rules may not apply to it, and it does not determine inclusion in search.
  • PerplexityBot: Perplexity search results. Perplexity says it is not used to crawl content for AI foundation models and recommends allowing it. Perplexity-User acts on user requests and, according to Perplexity, generally ignores robots.txt.

What to do: open yourdomain.com/robots.txt and check that none of the names listed above has a Disallow rule. If Gemini matters to you, do not block Google-Extended. If you do not want your content used to train OpenAI's models, you can block GPTBot and still be visible in ChatGPT search.

What Google says: GEO is still SEO

Google has a dedicated guide on optimizing for generative AI features in Search. The key line: from Google Search's perspective, optimizing for generative AI search is optimizing for search, and thus still SEO. Google also says there is no special schema.org markup you need to add. Beyond the usual technical requirements for Search there is one condition: the site must not be excluded from AI features in the Search Console settings, more on that below.

The same guide lists things you can skip: chunking text into small pieces for AI, rewriting content just for AI, chasing inauthentic mentions, and overfocusing on structured data. What matters most, according to Google, is unique content based on first-hand experience rather than restating what is already online. For a local business that is good news: nobody knows your services, prices and neighborhood better than you do.

New in 2026: Search Console has a Search generative AI setting under Settings. As of August 31, 2026 it is available to all websites. By default a site is included in AI features, but if someone excluded it, its links and content will not appear in AI Overviews, AI Mode or the AI features in Discover. What to do: check this setting once and make sure your site is not excluded.

llms.txt: an honest look at a popular file

llms.txt is a proposal by Jeremy Howard from September 2024: a Markdown file added to a website, with a short overview and links, meant to make the site's content easier for language models to use. The project's own site calls it a proposal open for community input. Some AI companies publish such files for their own developer documentation, but that is not a statement that their assistants read llms.txt on business websites.

Google is clear: Google Search does not use LLMS.txt files, and adding one will neither help nor hurt your visibility. In OpenAI's and Perplexity's crawler documentation I found no statement that their assistants use them. What to do: if you already have the file, you can keep it. If you don't, there are more important items on the list below.

What to do on your site and beyond: priorities for a local business

This is the order I work in for small businesses. For each item I note whether a vendor recommends it or whether it is reasonable practice without a vendor guarantee.

  • Access and indexing: site verified in Google Search Console and Bing Webmaster Tools, no blocks for Googlebot, OAI-SearchBot or PerplexityBot. Recommended by Google, OpenAI and Perplexity.
  • IndexNow: notifies search engines when URLs are added, updated or removed. Microsoft recommends it so that AI answers rely on the current version of a page.
  • Specific content that answers customers' real questions: services, service area, opening hours, prices or price ranges if the business publishes them, an FAQ section with real answers. Microsoft names clear headings, tables and FAQ sections, Google emphasizes unique, well-organized content.
  • Google Business Profile and Bing Places: current address, hours and phone number. Google recommends keeping Business Profile information up to date, Microsoft recommends registering with Bing Places for location-based queries.
  • Structured data, for example LocalBusiness and Service, matching what is visible on the page. Google: not required for AI features, but worth keeping as part of SEO. Practice without a guarantee of AI visibility.
  • A consistent name, address and phone number across directories and industry sites. Reasonable practice, no vendor describes it as a factor for assistants.
  • Genuine reviews and mentions on other sites. Google says its AI features can show what people say about products and services in blogs, videos and forums, but seeking inauthentic mentions does not help. Practice without a guarantee.

How to check whether it works

In 2026 the first free reports on visibility in AI answers became available.

  • Google Search Console, Generative AI performance report: impressions of links to your site in AI Overviews and AI Mode, broken down by page, country, device and date. Available to all websites since August 31, 2026. The report may not appear if your site has too few impressions in AI features.
  • Bing Webmaster Tools, AI Performance (public preview since February 10, 2026): number of citations in Copilot, in Bing's AI summaries and in select partner integrations, cited pages, and a sample of the phrases the AI used to retrieve your content.
  • ChatGPT and Perplexity: in the documentation I reviewed I found no comparable report for site owners. That leaves manual checks and watching your site analytics for visits coming from these services.

A manual test makes sense too: a few questions a customer would ask, checked once a month in several assistants. Answers are not fixed, though, and may look different for different people and at different times, so one screenshot proves nothing.

What nobody can promise

No vendor guarantees that an assistant will recommend a particular business. Google says outright that even a page meeting every requirement is not guaranteed to be indexed or shown, and warns against tools that promise ranking success or claim to use internal Google metrics. A GEO offer with a guaranteed spot in ChatGPT is a promise nobody can keep.

What you can do is remove the obstacles and give assistants a good source: an accessible site with specific information, consistent business details and genuine reviews. This is the cleanup I do as part of SEO and GEO work: I check robots.txt and indexing, connect Search Console, Bing Webmaster Tools and IndexNow, tidy up content and structured data, and then review the reports every month to see what changes.

Frequently asked questions

Do I need to add an llms.txt file?

No. Google says Google Search does not use it and that it neither helps nor hurts visibility. In OpenAI's and Perplexity's crawler documentation I found no statement that they read it. You can have one, but it is not a priority.

If I block GPTBot, will I disappear from ChatGPT?

According to OpenAI, no. GPTBot relates to content for model training, while OAI-SearchBot decides inclusion in ChatGPT search. OpenAI says the two settings are independent.

Is structured data required to appear in AI Overviews?

No. Google says there is no special schema.org markup required for AI features. It is worth having as part of regular SEO, but specific content on the page matters more.

How fast do robots.txt changes take effect?

OpenAI says it can take about 24 hours for ChatGPT search, and Perplexity says up to 24 hours. The effect of content changes depends on when crawlers revisit the site, which is why IndexNow and watching the reports over a longer period help.

monthly_tip.md

@ newsletter

One short tip a month

How to win clients through your website, Google and AI assistants. One email a month, no spam. Unsubscribe with a single click.

Sources

  1. Google Search Central: Optimizing your website for generative AI features on Google Search
  2. Google Search Console Help: Search generative AI control (2026)
  3. Google Search Console Help: Generative AI performance report (Search) (2026)
  4. Google for Developers: Google's common crawlers, Google-Extended
  5. OpenAI: Overview of OpenAI Crawlers
  6. Perplexity: Perplexity Crawlers
  7. Microsoft Bing Blog: Introducing AI Performance in Bing Webmaster Tools Public Preview (February 10, 2026)
  8. llmstxt.org: The /llms.txt file, v2 (J. Howard, September 3, 2024, modified August 10, 2026)

Related