Last updated · reviewed by the Essentras editorial team
AI visibility guide
How to get your business recommended by ChatGPT
No one outside OpenAI can promise a ChatGPT recommendation. ChatGPT search, Perplexity, Gemini and Claude answer from pages their search systems can reach, so allow their search crawlers (OAI-SearchBot for ChatGPT), stay indexed in Google and Bing, keep your business facts identical everywhere, and earn independent reviews and coverage that answers can cite.
OpenAI, Perplexity, Google, Anthropic and Microsoft documentation, read on that date
What you will do
- Check that each assistant's search crawler can reach your site, with copy-ready robots.txt rules.
- Get into the indexes answers draw on: Google, Bing with IndexNow, and OpenAI’s product feeds if you sell products.
- Make your business facts consistent: website, structured data, Google Business Profile and Bing Places.
- Build the third-party evidence assistants cite, without fake reviews or hidden text.
- Measure where you appear, for free, before you spend anything.
Start here in 1 click
Measure first. These 3 tools are free and show where you stand today. None of them, ours included, can make an assistant recommend you.
Free AI visibility score
Paste any URL for a score from 1 to 100 of how likely ChatGPT, Claude, Perplexity and Google AI Overviews are to cite that page, with a ranked fix list. No signup.
Free AI check
We look up what your buyers really search for, ask an AI those questions now and show the answer word for word, with the competitors it names instead of you.
AI Visibility Tracker
$0 to start
- Free plan
- 3 questions, asked every month
- Assistants
- 7, including ChatGPT, Gemini, Claude and Perplexity
- Pro
- $49 a month: 25 questions every week
Bing Webmaster Tools also reports, for free, how often Microsoft Copilot and Bing’s AI answers cite your pages. Step 5 below explains it.
How assistants find sources
How ChatGPT, Perplexity, Gemini, Claude and Copilot find sources, in their own words
Each company documents 1 thing clearly: which crawler feeds its search answers and how to let it in. None publishes a formula for which businesses it recommends. This table is what they do say.
| Assistant | Search crawler | Training control | Fetches for a user | What the docs say about appearing |
|---|---|---|---|---|
| ChatGPT (OpenAI) | OAI-SearchBot: surfaces sites in ChatGPT search | GPTBot: content that may be used to train OpenAI’s models | ChatGPT-User: visits pages when a user asks; robots.txt rules "may not apply" | Allow OAI-SearchBot and OpenAI’s published IP ranges. Blocked sites are not shown in ChatGPT search answers, except as navigational links. |
| Perplexity | PerplexityBot: surfaces and links sites in Perplexity search | PerplexityBot is "not used to crawl content for AI foundation models" | Perplexity-User: fetches pages for a question and "generally ignores robots.txt rules" | Allow PerplexityBot and its published IP ranges; a firewall may need an explicit allow rule. |
| Google AI Overviews and AI Mode | Googlebot, as for Search | Google-Extended token; it does not affect Search | Not applicable | Indexed and eligible for a snippet. No special optimization, files or schema needed. |
| Gemini apps | Google’s crawling; Google-Extended governs grounding in Gemini Apps | Google-Extended token | Not stated | Disallowing Google-Extended limits training and grounding in Gemini Apps, not Google Search. |
| Claude (Anthropic) | Claude-SearchBot: indexes content to improve search results | ClaudeBot: content that could contribute to training | Claude-User: fetches pages for a user’s question | Blocking Claude-SearchBot or Claude-User may reduce your visibility in Claude’s search answers. Claude’s web search tool always returns citations. |
| Microsoft Copilot | Bing’s crawler and index | Not covered in the pages we read | Not covered | Bing’s AI Performance report covers citations in Copilot and Bing’s AI summaries. Bing says it respects robots.txt. |
What none of them document: a list of businesses they recommend, a way to pay for a place in an answer, or a ranking formula. OpenAI does sell ads in ChatGPT through its own ads platform; buying an ad is a different product from being recommended in an answer.
Step 1
Step 1: let the search crawlers in
An assistant cannot cite a page its search crawler cannot fetch. The settings are independent, so you can allow search and still refuse training.
1. Open your robots.txt
Visit your domain followed by /robots.txt. Look for groups naming OAI-SearchBot, PerplexityBot, Claude-SearchBot or Googlebot, and for a "User-agent: *" group with "Disallow: /", which blocks every crawler.
2. Allow the search crawlers
Let OAI-SearchBot, PerplexityBot, Claude-SearchBot, Googlebot and Bing’s crawler fetch your public pages. These are the crawlers behind search answers.
3. Decide training separately
GPTBot and ClaudeBot govern training. Google-Extended governs training and grounding in Gemini Apps, so blocking it can also limit Gemini answers. None of them controls ChatGPT search or Google Search inclusion.
4. Check your firewall and CDN
Bot protection can block crawlers your robots.txt allows. Perplexity documents firewall allow rules, and Google lists crawling allowed "by any CDN or hosting infrastructure" among its basics.
5. Use the published IP ranges
OpenAI and Perplexity publish JSON files of their crawler IP addresses, so you can confirm a visitor really is their crawler before you allow it.
6. Give it a day, then test
OpenAI says changes take about 24 hours to reach ChatGPT search, Perplexity up to 24 hours. Then ask each assistant about the page and see whether it cites it.
User-triggered fetchers are different. ChatGPT-User and Perplexity-User act for a person asking a question, and both companies say robots.txt may not apply to them. Anthropic says disabling Claude-User prevents Claude from retrieving your content for user queries.
Step 2
Step 2: be in the indexes the answers draw on
Search-based answers are built from search indexes. Being in them, with current versions of your pages, comes before anything else.
Google’s AI Overviews and AI Mode show only pages that are indexed and eligible for a snippet. Search Console shows your indexing, and Google counts traffic from AI features in its Performance report.
Bing
Verify your site in Bing Webmaster Tools and submit an XML sitemap with accurate lastmod dates; Bing ignores changefreq and priority. Bing says freshness signals affect how fast updates reach its results and AI-generated answers.
IndexNow
A ping that tells participating search engines a URL was added, updated or deleted. Microsoft Bing, Naver, Seznam.cz, Yandex and Yep support it, and Bing says it helps AI systems use the current version of a page.
Product feeds for ChatGPT
Merchants can share a structured product feed with OpenAI under the Agentic Commerce Protocol. OpenAI says feeds give ChatGPT "the catalog data it needs to index your products": titles, descriptions, images, price and availability.
What OpenAI will not list
OpenAI only allows products that are legal, safe and suitable for a general audience. Alcohol, nicotine, gambling, weapons and prescription-only medicines are among the exclusions.
Local businesses
Keep your Google Business Profile verified and complete, and register with Bing Places for Business, which Microsoft names for keeping address, hours and contact details current for AI answers.
Step 3
Step 3: make your business facts identical everywhere
Assistants describe a business from many pages at once. When your name, address, category and claims match everywhere, there is nothing for them to reconcile.
- One name, written the same way on your site, Google Business Profile, Bing Places, social profiles and directories.
- An About page that says what you do, where, since when and for whom, in plain sentences an assistant can quote.
- Organization structured data on your home page. Google says it helps Google "better understand your organization's administrative details and disambiguate your organization in search results", and it has no required properties.
- sameAs links in that markup to your official profiles, and those profiles linking back to your site.
- Structured data that matches the visible text on the page: Google lists this among its basics for AI features.
- Merchant Center and Business Profile information kept up to date, which Google lists among the same basics.
- A Wikidata item only if your company meets Wikidata’s criteria, and a Wikipedia article only if it is notable. Both have strict rules for connected editors.
The Google side of a consistent entity, step by step: How to get a Google Knowledge Panel
Step 4
Step 4: earn the reviews and coverage that answers cite
OpenAI, Anthropic and Perplexity publish nothing about what makes a source citable beyond letting their crawlers in. Google and Microsoft say a little more, and both point the same way: independent, accurate, well-structured information.
Reviews, for local answers
Google says local results are mainly based on relevance, distance and prominence, and that "More reviews and positive ratings can help your business's local ranking." Ask every real customer.
What Google bans
Google Maps removes reviews paid for "directly or in kind" and reviews bought with discounts or free goods, and bans asking only happy customers to review.
Independent coverage
Coverage a journalist chose to write is the strongest third-party evidence of what you do. Labeled paid placements exist, but they buy publication, not a mention in any answer.
Pages that are easy to cite
Bing advises clear headings, tables and FAQ sections, examples and data with sources, regular updates, and text, images and video that describe the same products consistently.
Listings in your category
Keep the facts about you current on the comparison and review sites your buyers use. Correct errors through each site’s own process, never with fake accounts.
Wikipedia, only if notable
A Wikipedia article is not a marketing channel, and paid placements never count toward it. If your company qualifies, follow its disclosure rules.
Whether your company qualifies for Wikipedia, and how to disclose: How to get a Wikipedia page for your company
llms.txt
llms.txt: a proposal, not a ranking switch
llms.txt is a proposed file at the root of a website that gives AI agents a short markdown guide to its content. Jeremy Howard proposed it on 3 Sep 2024; version 2 is dated 10 Aug 2026.
| Question | What the sources say |
|---|---|
| What it is | "A proposal to standardise on using an /llms.txt file to provide information to help agents use a website": a markdown file with an H1, a short summary and lists of links to markdown pages. |
| Who publishes one | The proposal says thousands of sites do, including OpenAI, Anthropic and Google for their own developer docs. We saw such indexes on OpenAI’s and Perplexity’s documentation sites. Essentras publishes one too. |
| Google AI features | "You don't need to create new machine readable files, AI text files, or markup to appear in these features." |
| OpenAI, Anthropic, Perplexity and Microsoft | None of the crawler and search pages we read says llms.txt affects whether you appear or are recommended. |
| Verdict | Cheap to add and useful for agents reading your documentation. Not a lever for recommendations. |
Step 5
Step 5: measure before and after, for free
Answers change by question, by assistant and over time. Ask the same questions on a schedule and compare the answers word for word.
Bing AI Performance
Free in Bing Webmaster Tools since its public preview of 10 Feb 2026: citations of your pages in Microsoft Copilot, AI-generated summaries in Bing and select partner integrations, the pages cited and the grounding queries behind them.
Topics Portfolio
A beta tab added on 13 Oct 2026 groups those grounding queries into topics and lists queries where other sites are cited and yours is not. Bing calls its numbers directional.
Google Search Console
Google counts clicks and impressions from AI Overviews and AI Mode inside the regular Performance report, and suggests tracking conversions and time on site in tools such as Google Analytics.
Referral visits
In your analytics, look at visits referred by the assistants’ own domains, such as chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com and claude.ai, and at the pages they land on.
A fixed question set
Write down the questions your buyers ask, ask each assistant every month and keep the answers. Our free AI check and tracker do this across assistants and keep every answer.
What to record every time
- The date, the assistant, and whether web search was on.
- The question, word for word, as a buyer would ask it.
- Whether you were named, and where in the answer.
- Who was named instead.
- Which sources were cited.
- Anything wrong or outdated about you, and where it came from.
What does not work
What does not work, and what breaks the rules
Most of what is sold as a shortcut is either impossible or against a published policy.
Buying a promised recommendation
No assistant on this page offers one, and no outside vendor can deliver it. OpenAI sells ads, which are a different product.
Hidden text
Text hidden from visitors to manipulate search systems breaks Google’s spam policies, which name white text on a white background, text behind an image and a font size of 0.
Mass-produced pages
Generating many pages "for the primary purpose of manipulating search rankings" is scaled content abuse under Google’s spam policies, however the pages are made.
Fake or incentivized reviews
Google Maps removes reviews paid for in cash or in kind, and bans offering anything in exchange for a review.
Blocking the search crawlers
Blocking OAI-SearchBot keeps your pages out of ChatGPT search answers; blocking Claude-SearchBot may reduce your visibility in Claude’s results.
Treating llms.txt or schema as a switch
Google says no special files or schema are needed for its AI features, and the other companies document none.
Optimizing for 1 assistant
Each uses its own crawler and index. Measure every assistant your buyers use.
A paid Wikipedia article
Wikipedia needs independent coverage, and undisclosed paid editing breaks its terms. It is not a shortcut into AI answers.
Copy-ready
Copy-ready templates
Replace everything in [brackets]. Test any robots.txt change on a staging copy first, and keep the rules your site already needs.
robots.txt: allow AI search, refuse training
You want to be citable in search answers but keep your content out of model training.
# AI search crawlers: allowed, so answers can cite these pages User-agent: OAI-SearchBot Allow: / User-agent: PerplexityBot Allow: / User-agent: Claude-SearchBot Allow: / # Model training: refused User-agent: GPTBot Disallow: / User-agent: ClaudeBot Disallow: / # Google-Extended also controls grounding in Gemini Apps. # Allowed here; disallow it only if you accept less Gemini grounding. User-agent: Google-Extended Allow: /
Organization structured data (JSON-LD)
On your home page. Every value must match what the page shows.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "[Your business name, exactly as on your site]",
"url": "https://[yourdomain]",
"logo": "https://[yourdomain]/[logo file]",
"description": "[1 plain sentence: what you do, for whom, where]",
"email": "[contact address]",
"telephone": "[+country code and number]",
"address": {
"@type": "PostalAddress",
"streetAddress": "[street]",
"addressLocality": "[city]",
"postalCode": "[postcode]",
"addressCountry": "[2-letter country code]"
},
"sameAs": [
"https://[official profile 1]",
"https://[official profile 2]"
]
}
</script>AI answer log
Fill in 1 entry per question, per assistant, every month.
Date: [date] Assistant: [ChatGPT / Gemini / Claude / Perplexity / Copilot] Web search on? [yes / no] Question: [exactly as a buyer would ask it] Named us? [yes / no] Where: [first / later / not named] Named instead: [competitors] Sources cited: [URLs] Wrong or outdated facts about us: [list, with the source they came from] Action: [fix our page / correct a listing / none] Owner: [name]
Correction request to a site that has your facts wrong
A directory, review site or article lists an old address, a closed service or a wrong fact.
Subject: Correction request: [page title] Hello, Your page [URL] lists [business name] with [the wrong fact, for example our old address]. The correct information is [correct fact], as shown on [our official page URL or an independent source]. Could you update it? I am happy to answer any questions. [Name], [role], [business name] [Contact details]
Review request that follows Google’s rules
Send to every customer, not only happy ones, with no discount or gift in exchange.
Subject: How did we do? Hi [first name], Thank you for choosing [business name] for [service]. If you have a minute, we would value an honest review, good or bad: [review link] Thank you, [Name]
Optional help
Optional help, and free starts
Everything above can be done in-house. If you want help, Essentras tracks what AI assistants say about you and works on the sources they already quote in your category. No one can sell a position inside a model, and we do not. The step-by-step method we use is public: the LLM SEO playbook.
Optional. You can follow every step on this page without buying anything.
Often added
AI Visibility Tracker Pro
$49 a month, plus $29 per extra domain
Your own 25 questions every week across 7 assistants, every answer kept word for word with its sources, and a Monday report.
See the trackerAI SEO
Free assessment, then a quote
Work on the sources AI assistants already quote in your category. No position in any model is promised.
See how it worksBusiness Knowledge Panel
$1,200
One consistent entity for your brand on Google. Placements are never required for it.
See the panel serviceLabeled placements on named outlets
Priced per outlet
Publication on a named outlet, with the label each listing states. It buys a published page, not a mention in any AI answer.
Browse outlets with prices
Not ready yet?
Free AI visibility score
$0
A 1 to 100 score for 1 page, with a ranked fix list.
Score a pageFree AI check
$0
Real buyer questions, the answer word for word, and the competitors named instead.
Run the checkAI Visibility Baseline
$97
A one-time read of what AI assistants already know about you, and the first fixes.
See the baseline
FAQ
Frequently asked questions
Can I pay ChatGPT to recommend my business?▾
No. OpenAI documents no way to pay for a place in ChatGPT’s answers. It does sell ads in ChatGPT through its ads platform, which is a separate product. Anyone selling a promised recommendation is selling something OpenAI does not offer.
How does ChatGPT decide which businesses to recommend?▾
OpenAI does not publish a formula. Its documentation says ChatGPT search surfaces websites its OAI-SearchBot crawler can reach. What you can work on is access, accurate and consistent facts, and independent reviews and coverage.
Does ChatGPT use Google or Bing?▾
OpenAI’s crawler documentation describes its own search crawler, OAI-SearchBot, and names no search partner. OpenAI’s help center did not load for our reader on 19 Oct 2026. Microsoft reports Copilot citations of your pages in Bing Webmaster Tools, and Google’s AI features use Google’s index.
Should I block GPTBot?▾
It is your choice. OpenAI says GPTBot is for training and that each setting is independent: you can block GPTBot and still allow OAI-SearchBot, the crawler that decides whether your pages can appear in ChatGPT search.
Does llms.txt help ChatGPT recommend my business?▾
No official statement says it does. llms.txt is a proposal for helping agents read a website, and Google says no AI text files are needed for its AI features. It is cheap to add, but it is not a lever for recommendations.
Do reviews affect ChatGPT recommendations?▾
OpenAI does not say. Google says more reviews and positive ratings can help a business’s local ranking. Ask every real customer for an honest review, and never pay for reviews or offer anything in exchange.
How long does it take for ChatGPT to notice changes?▾
OpenAI says robots.txt changes take about 24 hours to reach ChatGPT search. Content changes depend on when your pages are fetched again; for Google, crawling can take from several days to several months.
How do I know if ChatGPT recommends my business?▾
Ask it the questions your buyers ask, in a fresh session, and record the answers and sources. Our free AI check does this from real search data, and Bing Webmaster Tools reports Copilot citations of your pages for free.
Does schema markup get me into AI answers?▾
Google says no special schema is needed for AI Overviews and AI Mode, but structured data should match the visible text, and Organization markup helps Google understand and disambiguate your business. It is housekeeping, not a switch.
Next steps
- 01
Measure today
Run the free AI check or score and note what each assistant says and cites.
- 02
Open the doors
Allow OAI-SearchBot, PerplexityBot and Claude-SearchBot; decide training crawlers separately.
- 03
Fix your facts
The same name and details everywhere, and Organization markup that matches the page.
- 04
Earn evidence
Real reviews, independent coverage and current listings on the sites your category relies on.
- 05
Measure again every month
The same questions and assistants, with the answers kept word for word.
See what AI says about you today
Free and live: real buyer questions, the answer word for word, and the competitors named instead of you.
Run the free AI check Or score a single pageChecked on . Every rule, price and figure on this page comes from the official pages listed below, read on that date, or from the Essentras price list. Rules change: if a page below disagrees with this guide, the page wins.
Sources
- OpenAI: Overview of OpenAI crawlers, read 19 Oct 2026
- OpenAI: Commerce, get started with product feeds, read 19 Oct 2026
- OpenAI: Ads documentation, read 19 Oct 2026
- Perplexity: Perplexity Crawlers, read 19 Oct 2026
- Google Search Central: AI features and your website (last updated 10 Dec 2025), read 19 Oct 2026
- Google Search Central: Google’s common crawlers, Google-Extended (last updated 15 Oct 2026), read 19 Oct 2026
- Google Search Central: Organization structured data (last updated 8 Sep 2026), read 19 Oct 2026
- Google Search Central: Spam policies (last updated 28 Aug 2026), read 19 Oct 2026
- Google Business Profile Help: Tips to improve your local ranking on Google, read 19 Oct 2026
- Google Maps: user contributed content policy, fake engagement, read 19 Oct 2026
- Anthropic: Does Anthropic crawl data from the web? (7 Apr 2026), read 19 Oct 2026
- Claude docs: Web search tool, read 19 Oct 2026
- Bing Webmaster Blog: Introducing AI Performance in Bing Webmaster Tools (10 Feb 2026), read 19 Oct 2026
- Bing Webmaster Blog: Introducing Topics Portfolio (Beta) in AI Performance (13 Oct 2026), read 19 Oct 2026
- Bing Webmaster Blog: Keeping content discoverable with sitemaps in AI powered search (31 Jul 2025), read 19 Oct 2026
- IndexNow, read 19 Oct 2026
- The /llms.txt file, v2, by Jeremy Howard (published 3 Sep 2024, modified 10 Aug 2026), read 19 Oct 2026
Essentras sells AI visibility tracking, AI SEO work and media placements. No assistant’s recommendation can be bought or promised, by us or anyone. Company documentation changes often: check the pages linked above before acting.
More on AI visibility
Adds us as a preferred source, so our guides surface for you in Search and AI answers. One click, and you can undo it any time.
Free tool · Authority Audit
Where does your authority actually stand?
Before you fix one piece, see the whole picture. A free Authority Audit scores your presence across Google, AI, Wikipedia, reviews, and press, and shows you exactly where to start.