To get your business cited by ChatGPT, allow OpenAI’s OAI-SearchBot in robots.txt, keep your key facts in plain text on clear service pages, describe the business the same way everywhere, and earn mentions on trusted third-party sites. Nobody can buy or guarantee a citation; you can only make your business easy to find and confirm.
A growing number of buyers now ask ChatGPT the question they used to type into Google: who does this kind of work, which company fits my situation, what should I ask before I hire. When the answer names three businesses and yours is not one of them, it feels like a door closed without anyone telling you why.
The good news is that the reasons are rarely mysterious. In most cases the business is missing from AI answers for ordinary causes: the site blocks the crawler, the important facts sit inside images or vague marketing lines, the company is described differently on every profile, or almost nobody outside the business has written about it. This guide covers each cause and what to do about it, in the order we work through it on real projects. It also covers what to skip, because this topic attracts a lot of tactics that cost money and change nothing.
One honest note before we start. OpenAI does not publish a ranking formula for ChatGPT, and no outside company has access to one. What follows is based on what OpenAI and Google document publicly, plus the long-standing basics of being findable on the web. Where something is an educated practice rather than a documented rule, we say so.
How does ChatGPT decide which businesses to mention?
ChatGPT can produce an answer about businesses in two different ways, and the difference matters for what you can influence.
Answers from what the model already learned
A language model learns patterns from a large body of text during training. If your business was written about often and consistently before that training happened, the model may “know” it and mention it without looking anything up. You cannot edit that memory directly, and you cannot speed it up. What you can do is make sure that whatever is written about you across the web is accurate and consistent, so future training data describes you correctly.
Answers from live web search
The second path is search. When ChatGPT searches the web for a question, it reads pages and shows the sources it used as links next to the answer. This is where a citation, in the strict sense, comes from: a clickable reference to a page. OpenAI states that OAI-SearchBot is used to surface websites in search results in ChatGPT’s search features, and that sites opted out of it will not be shown in ChatGPT search answers.
That one sentence explains why so many businesses never appear: if your server, firewall or security plugin turns that crawler away, your pages cannot be one of the cited sources, no matter how good they are.
Mention versus citation
It helps to separate two outcomes. A mention is your business name appearing in the answer text. A citation is a link to one of your pages shown as a source. You can be mentioned because a directory or a news article named you, while the cited link points to that third-party page instead of your site. Both are valuable. A mention builds recognition with the person reading; a citation to your own page brings a visit you can measure.
| Path | Where the information comes from | What you can influence |
|---|---|---|
| Model knowledge | Text the model learned during training | Accurate, consistent coverage of your business across the web over time |
| Live search, your page cited | Pages OAI-SearchBot can access and that answer the question well | Crawler access, clear answer pages, plain-text facts |
| Live search, third-party page cited | Directories, reviews, articles and lists that name you | Profiles, reviews, press and partner mentions |
| Page visited during a chat | A user pastes your link or asks ChatGPT to open it | A fast page with the facts in readable text |
Can ChatGPT actually read your website?
This is the first thing to check, because every other step depends on it. OpenAI runs separate automated agents, and they are controlled separately in your robots.txt file:
- OAI-SearchBot crawls pages so they can appear in ChatGPT search answers. This is the one that decides whether your pages can be cited.
- GPTBot crawls content that may be used to train OpenAI’s models. Blocking it is a choice about training, not about search.
- ChatGPT-User visits a page when a person asks ChatGPT to open it during a conversation. OpenAI notes it is not used to decide what appears in search.
OpenAI’s documentation says these settings are independent, so a business can allow OAI-SearchBot and still disallow GPTBot. It also says a robots.txt change can take about 24 hours to take effect in its systems. If you want to be cited, the practical rule is simple: do not block OAI-SearchBot.
How to check your site in ten minutes
- Open yourdomain.com/robots.txt in a browser. Look for any line that names OAI-SearchBot, GPTBot or a wildcard rule such as “User-agent: *” followed by “Disallow: /”.
- Check your security layer. Firewalls, CDNs and WordPress security plugins can block bots before robots.txt is even read. Look for a “block AI crawlers” or “bot fight” setting and see what it covers.
- Check the important pages for noindex. A page marked noindex is telling crawlers to keep it out of results. Your service pages should not carry it.
- View the page source of a service page. If the main text is visible in the HTML, crawlers can read it. If the source is almost empty and the words only appear after scripts run, some crawlers may see much less than a visitor does.
- Ask your developer or host for server logs. Visits from OAI-SearchBot in the logs confirm it can reach the site. OpenAI publishes its crawler IP ranges, so a real visit can be told apart from a bot pretending to be one.
We see blocked crawlers more often than you would expect, usually because someone turned on a broad “block all bots” switch to reduce server load or spam and never looked at what it caught. If that happened on your site, fixing it is the single highest-value step in this guide.
Keeping a page out on purpose
The reverse also matters. OpenAI’s publishers and developers FAQ explains that a disallowed page can still appear as a bare link if it is found elsewhere, and recommends a noindex meta tag for pages you want kept out. The crawler has to be able to load the page to see that tag, so do not block and noindex the same page if your goal is removal.
How to get your business cited by ChatGPT: the plan, step by step
Once the crawler can get in, the work splits into five parts. They build on each other, and skipping the early ones makes the later ones much less effective.
| Step | What you do | Why it matters |
|---|---|---|
| 1. Access | Allow OAI-SearchBot, remove accidental blocks, keep money pages indexable | A page the crawler cannot read cannot be cited |
| 2. Identity | One name, address, phone and description everywhere; matching structured data | Lets an AI system connect every mention to the same business |
| 3. Answers | One clear page per service, with direct answers to the questions buyers ask | Gives search a passage that fits the question |
| 4. Proof | Specific, verifiable facts and examples in plain text | Quotable statements are easier to cite than adjectives |
| 5. Mentions | Profiles, reviews, associations, press and partners that name you | Independent sources confirm what your own site claims |
None of these steps is a trick. They are the same things that make a business easy to find in Google and easy to trust for a human reader. That overlap is useful: work done for AI answers also helps classic search, and the reverse is true too. If the vocabulary is new, our guide to AEO vs SEO vs GEO and what each one does explains how the three disciplines fit together.
ChatGPT can only cite a business it can read, understand and confirm. Access, clarity and outside proof are the whole job; everything else is detail.
Make your business unmistakable
Search systems, AI or not, think in terms of entities: a specific business, with a name, a location, services and people, that can be connected to every page that mentions it. When your business is described five different ways across the web, that connection gets weaker. When the same facts repeat everywhere, it gets stronger.
Decide on one version of the facts
Write down, once, the facts that define the business: the exact legal or trading name, the address (or service area if you do not receive visitors), the main phone number, the year you started, the services you sell and the area you serve. Then make every place that describes you match that list word for word. That includes your website footer, your About page, your Google Business Profile, your Bing Places listing, social profiles, industry directories and any association member pages.
Small differences matter more than people expect. “Suite 207” on one profile and no suite on another, an old phone number on a forgotten directory, or a former business name on a social account can all make it harder for a system to be sure two mentions describe the same company.
Put an About page and a facts block on the site
Many small business sites have an About page that tells a story but never states the plain facts. Keep the story, and add a short, factual block: who you are, what you do, where you work, since when, and how to reach you. This gives any reader, human or machine, one page that answers “what is this business?” without interpretation.
Add structured data that matches the visible page
Structured data is code (usually JSON-LD) that describes the page in a standard vocabulary from schema.org. For a business, the usual types are Organization or LocalBusiness, with the name, address, phone, logo, and links to your real profiles in a “sameAs” list. Our beginner’s guide to schema markup covers the basics.
Two cautions. First, structured data should describe what is already visible on the page; Google explicitly recommends keeping it consistent with the visible text. Second, it helps machines understand the page, but it is not a secret switch for AI answers. Google’s own guidance says no special schema is required to appear in its AI features. Treat it as clarity, not as a shortcut.
Write pages that answer the questions buyers ask
When ChatGPT searches on someone’s behalf, it is looking for passages that answer a specific question. A home page that says “We deliver excellence across a range of solutions” answers nothing. A service page that says “We install and repair residential HVAC systems in single-family homes, with same-week appointments for repairs” answers several questions at once.
One service, one page
If you sell four services and they all live on one long “Services” page, each one gets a paragraph and none gets a real answer. Give every main service its own page with a clear title, a first paragraph that says exactly what the service is and who it is for, and sections for the questions customers really ask: what it involves, how the process works, what affects the cost, how long it usually takes in your experience, and what the next step is.
Lead with the answer
Put the direct answer at the top of each section, then the explanation. This pattern, sometimes called answer-first writing, helps a reader who skims and gives search systems a short, self-contained passage that matches the question. A good test: if someone read only the first two sentences of a section, would they have the answer?
Use the questions as headings
Headings written as real questions (“How much does a roof inspection cost?”, “Do you work with commercial properties?”) make the structure obvious. Use the words your customers use, not internal jargon. The best source for those questions is your own inbox and phone log: the questions people ask before they hire you are the questions they also ask ChatGPT.
Write for a person first
Pages stuffed with “best company near me” phrases or written by pasting AI output at scale tend to read badly and say little. Google’s guidance for its AI features is to focus on helpful, reliable content made for people, and there is no reason to expect ChatGPT search to reward the opposite. Clear, specific, useful pages are the goal for both.
A page also has to earn the visit once someone clicks the citation. If a page gets traffic but no inquiries, the problem has moved from visibility to conversion, which we cover in our guide on why a website is not getting leads.
Publish facts worth quoting
An AI answer has to state something. It is far easier to cite a page that states a clear, specific fact than one that only uses adjectives. “Award-winning, trusted, innovative” gives a system nothing to repeat. “Family-owned since 2009, licensed in three states, specializing in historic home restoration” gives it three facts it can use, as long as they are true.
What a quotable sentence looks like
- It names the subject. “Our company” is weaker than the business name, because a quoted sentence often travels without its surrounding context.
- It contains a checkable fact. A year, a service, a place, a process, a credential you actually hold.
- It stands alone. It still makes sense if it is the only sentence someone reads.
- It is true and current. If a number changes, update it everywhere at once.
On our own site, for example, we state that GoSocial has worked across web development and digital marketing since 2016, and that the domain, hosting account and code of a client site belong to the client. Those are short, checkable statements, and they match what we say on every page and profile.
Show your work
Case studies, project examples, before-and-after explanations, and clear descriptions of your process all give an AI system (and a buyer) something concrete. Describe what the client needed, what you did, and what changed, using only results you can stand behind. When a claim depends on outside data, cite the source. A page that links to the official rule or the original study is more trustworthy than one that just asserts it.
Never inflate
It is tempting to round up, add “leading” or claim results you cannot prove. Inflated claims are a risk twice: a person who checks them stops trusting you, and an AI system that finds contradictions between your site and other sources has a reason to prefer someone else. Accuracy is a competitive advantage here.
Earn mentions on sites ChatGPT already trusts
Your website is your own account of your business. Independent sources are what confirm it. When ChatGPT searches for “who does X in Y”, the pages it finds are often directories, review platforms, association listings, local news and comparison articles, not only company sites. If your business is absent from those pages, it is absent from the material the answer is built from.
Where mentions usually come from
- Business profiles. Google Business Profile, Bing Places, Apple Business Connect and the main directories in your industry. Complete them with the same facts as your site, real photos and accurate categories.
- Reviews. Genuine reviews on the platforms your customers use. Ask happy clients directly and make it easy. Never buy, write or incentivize fake reviews; the Federal Trade Commission’s final rule banning fake reviews and testimonials makes that a legal risk as well as a trust risk.
- Associations and licensing bodies. Member directories, certification lists and state license lookups are exactly the kind of independent confirmation a system can trust.
- Partners and suppliers. Manufacturers, platforms and software vendors often list certified installers or partners. Ask to be listed.
- Local and trade press. A quote in a trade publication or a local news story about a real project names you on a site with its own reputation.
- Podcasts, webinars and community talks. Episode pages and event listings usually name the guest and the company.
For service businesses that work in a defined area, profile and review work overlaps heavily with local search. Our guide to local SEO and ranking higher in your area goes deeper on profiles, categories and reviews.
What not to do
Avoid paid link schemes, mass directory submissions to sites nobody reads, and “AI visibility” packages that promise placement. There is no submission form for ChatGPT answers and no paid placement inside them. Anyone selling a guaranteed AI recommendation is selling something they do not control.
Get the technical foundations right
Even with the crawler allowed, a page has to be readable and reachable. These are the technical basics that most often get in the way, and they are the same ones that matter for classic search.
Keep the important content in text
Services listed only inside a banner image, prices only in a downloadable PDF, or the main description loaded by a script after the page opens are all harder for crawlers to read reliably. Google lists “keeping important content in text form” among its recommendations for AI features. If a fact matters, it should be a sentence in the HTML.
Link your pages together
Crawlers discover pages by following links. A service page that is only reachable from a dropdown menu, or a case study that no other page links to, is harder to find and seems less important. Link related pages to each other with descriptive anchor text, and keep an up-to-date XML sitemap.
Make pages fast and stable
Speed is not a direct “AI ranking factor” that anyone can point to, but a slow or broken page wastes the visit when someone clicks the citation, and a server that times out under crawler load may not get read at all. Our explainer on Core Web Vitals and how they affect SEO covers the metrics Google measures.
Fix the platform if the platform is the problem
Some sites cannot be fixed page by page: the theme hides text in sliders, the builder outputs pages that load slowly everywhere, or the owner cannot edit the content without a developer. In that case a rebuild on a cleaner foundation may be the better investment. Our comparison of a custom website vs WordPress explains how to choose, and our web development service describes how we build sites for search and AI readability from the start.
If you want someone to look at these foundations with you, a Blueprint Briefing is a short call where we go through crawler access, page structure and what your site has to produce. You leave with a list, whether or not you work with us.
Do llms.txt files or special markup help?
You will find many articles telling you to add an llms.txt file, special “AI schema” or a dedicated machine-readable summary to get into AI answers. The evidence is thinner than the marketing suggests.
What the platforms actually say
Google is direct about its own AI features. Its guidance on AI features and your website says there are no additional requirements to appear in AI Overviews or AI Mode, no special schema.org structured data you need to add, and no need to create new machine-readable files or AI text files. OpenAI’s crawler documentation explains robots.txt controls and does not ask site owners for an llms.txt file.
So what is llms.txt?
llms.txt is a proposed convention: a plain text or Markdown file at the root of a site that summarizes the most useful pages for language models. Some sites publish one, and it does no harm. But it is optional, it does not replace robots.txt, and there is no public evidence from OpenAI or Google that it changes who gets cited. If a vendor makes it the center of an AI visibility plan, ask what else they plan to do.
Where structured data still earns its place
Structured data is still worth doing well, for the reasons covered earlier: it describes your business precisely, it supports rich results in classic search where Google still shows them, and it reduces ambiguity about who you are. Just do not expect it to compensate for a blocked crawler, thin pages or a business nobody else mentions.
How do you measure whether ChatGPT cites you?
Measurement here is less precise than in Google Search, where Search Console reports impressions and clicks. There is no equivalent report for ChatGPT answers. You can still build a reliable picture from a few sources.
Referral traffic in your analytics
According to OpenAI’s publishers FAQ, ChatGPT adds the parameter utm_source=chatgpt.com to links in its search results. In Google Analytics 4, that traffic appears under the session source chatgpt.com. Create a simple exploration or a saved report filtered to that source and watch which landing pages it reaches. Those pages are the ones being cited. If GA4 is new to you, our guide to tracking website analytics to improve SEO covers the basic reports.
A fixed set of test questions
Write down ten to fifteen questions a real buyer would ask, in their words: “Who installs solar panels for small businesses in my state?”, “What should I ask before hiring a bookkeeper?” Run them in ChatGPT once a month, note whether your business is mentioned, which sources are cited and how you are described. Answers vary from one run to the next, so look for patterns over several months rather than reacting to a single answer.
Server logs
Logs show whether OAI-SearchBot and ChatGPT-User are visiting and which pages they request. A crawler that never reaches your service pages is a discovery problem; a crawler that reaches them while you never get cited points to content or authority.
Leads that mention AI
Add “ChatGPT or another AI assistant” as an option in your “How did you hear about us?” question, and ask on the first call. It is the simplest measure and often the most revealing one.
How long does it take to get cited?
There is no honest fixed answer, and you should be wary of anyone who gives you a number of days. The one timing OpenAI documents is that a robots.txt change can take about 24 hours to be reflected in its systems. Everything after that depends on factors nobody controls fully: when pages are crawled again, how competitive the question is, how many trusted sources already answer it, and how quickly other sites start mentioning you.
In practice the work moves at different speeds. Removing a crawler block is fast. Rewriting service pages takes a few weeks of real effort. Building mentions, reviews and press is slow and ongoing, and it compounds: each independent source makes the next answer more likely to include you. Treat this as part of running the business, not as a one-time project.
No one can guarantee a citation in ChatGPT. What a business can guarantee is that its site is readable, its facts are clear and consistent, and that real people and organizations talk about it.
We hold ourselves to that line. We do not promise a citation or a ranking on any project; we build for the things that make one more likely and show you the measurements.
What mistakes keep businesses out of AI answers?
These are the problems we find most often when a business asks why it never shows up. Most are easy to fix once someone looks.
- Blocking the crawler by accident. A “block AI bots” switch in a CDN or security plugin, or a robots.txt copied from another site, quietly removes the business from ChatGPT search answers.
- Hiding facts in images and PDFs. The service list lives in a graphic, the price sheet in a PDF, the address in a map embed. A crawler reads text best.
- Vague home and service pages. Slogans instead of statements. If the page does not say what you do, for whom and where, nothing can quote it.
- Inconsistent business details. Different names, old addresses and dead phone numbers across profiles make it harder to connect every mention to one business.
- No outside mentions. A business that exists only on its own website has nothing independent to confirm it.
- Mass-produced pages. Hundreds of near-identical city pages or AI-written articles with no real knowledge behind them add little and can drag down trust in the whole site.
- Buying promises. Paying for “guaranteed AI placement” or “ChatGPT ranking” packages. The money is better spent on the steps above.
If several of these sound familiar, start with the first two. They are the cheapest to fix and the most likely to be stopping everything else from working.
A checklist to start this week
If you only have a few hours, work through this list in order. Each item is something you or your web team can do without special tools.
- Check robots.txt and your security settings for anything that blocks OAI-SearchBot, and remove accidental blocks.
- Confirm your service pages are indexable and that their main text is visible in the page source.
- Write your one-version fact list: name, address or service area, phone, year founded, services, area served.
- Update every profile to match it: Google Business Profile, Bing Places, social accounts and the main directories in your industry.
- Add or fix a facts block on your About page and make sure your Organization or LocalBusiness structured data matches it.
- Pick your most important service page and rewrite its first paragraph so it states exactly what you do, for whom and where.
- Add the five questions customers ask most to that page, each answered in the first two sentences under its heading.
- Ask three recent happy clients for a review on the platform they already use.
- Set up the chatgpt.com referral report in GA4 and a monthly list of test questions.
- List three outside places where your business should be mentioned and is not yet: an association, a supplier partner page, a trade publication.
Then repeat the page work for the next service, and keep the mention work going every month. This is the same process we follow on our own SEO, AEO and GEO service, where search visibility, answer-ready pages and AI citations are planned on one foundation.
Getting cited by ChatGPT is not a hack and not a mystery. It is the result of a site that machines can read, facts that are clear enough to quote, and a reputation that exists beyond your own pages. Do those three things well and you give every AI system, and every buyer, the best possible reason to name your business. You can see the kind of sites we build for that in our web design portfolio.
How to get your business cited by ChatGPT: FAQ
Can I pay to get my business recommended by ChatGPT?
No. There is no submission form and no paid placement inside ChatGPT answers. Businesses appear because ChatGPT can read their pages through search or because trusted sources describe them. Be cautious with any service that promises guaranteed placement in AI answers, because nobody outside OpenAI controls what it cites.
Does ChatGPT use my website to answer questions?
It can. When ChatGPT searches the web, OpenAI’s OAI-SearchBot is the crawler that surfaces websites in its search answers. If your robots.txt or firewall blocks that crawler, your pages will not be shown in ChatGPT search answers. A user can also paste your link and ask ChatGPT to read it.
Should I block GPTBot to protect my content?
That is a separate decision. GPTBot collects content that may be used to train OpenAI models, while OAI-SearchBot is the crawler used for search answers. OpenAI treats the two settings independently, so you can block GPTBot and still allow OAI-SearchBot if you want to be cited in search without contributing to training.
Do I need an llms.txt file to appear in ChatGPT?
No public OpenAI documentation requires one. llms.txt is an optional proposed file that summarizes a site for language models. It does no harm, but it does not replace robots.txt, clear pages or outside mentions. Google states that no special AI text files are needed to appear in its AI features.
How can I tell if ChatGPT is sending visitors to my site?
ChatGPT adds utm_source=chatgpt.com to the links in its search results, so the visits show up in Google Analytics 4 with chatgpt.com as the source. Check which landing pages receive that traffic, run a fixed set of buyer questions in ChatGPT each month, and ask new leads how they found you.
Does schema markup make ChatGPT cite my business?
Not on its own. Structured data such as Organization or LocalBusiness helps machines understand who you are, and it should match the visible page. But neither OpenAI nor Google presents it as a requirement for AI answers. Crawler access, clear pages and independent mentions matter more.
Is getting cited by ChatGPT different from ranking in Google?
The foundations overlap: crawlable pages, clear answers, trustworthy facts and outside mentions help both. The differences are in the crawler (OAI-SearchBot instead of Googlebot), the format (an answer with a few cited sources instead of a list of results) and the measurement, which relies on referral traffic and test questions rather than a dedicated report.

