There is no ChatGPT listing to claim and no AI directory to submit to. What an assistant says about your business comes from what it can read: your own site, your business listings, your reviews, and pages elsewhere that mention you. OpenAI and Google document a small part of how that works. Most of the advice written about it fills in the rest with guesses.
There is no listing to claim
Start with the vocabulary, because it is designed to confuse you. Answer engine optimization, generative engine optimization, LLM SEO, AI SEO: these are four names for the same work, sold by four sets of people. Nothing in any of them is a separate system you buy into. They all describe making your business legible to software that reads the web.
Here is what OpenAI actually publishes: a list of its crawlers, each doing a different job. OAI-SearchBot is “used to surface websites in search results in ChatGPT’s search features”. GPTBot is “used to make our generative AI foundation models more useful and safe” and crawls content “that may be used in training”. ChatGPT-User is “used for certain user actions in ChatGPT and Custom GPTs” — when someone asks a question, it “may visit a web page” then and there.
That is the entire documented surface. OpenAI publishes no ranking factors, no quality guidelines for businesses, and no directory. Its crawler documentation does not name the search index or the business-data provider behind a local answer, so when someone tells you ChatGPT recommends businesses because of some specific signal, ask them where that is written down.
What is documented and what is guesswork
Every row below is either something a provider states in its own documentation or something the internet decided on its own.
| Claim you will hear | Status | What the documentation says |
|---|---|---|
| ChatGPT crawls sites, and you control it in robots.txt | Documented | OpenAI lists OAI-SearchBot, GPTBot, ChatGPT-User and OAI-AdsBot, each with its own robots.txt token, and says changes can take “~24 hours” to take effect |
| Blocking GPTBot keeps you out of ChatGPT’s answers | Half right | GPTBot is the training crawler. OAI-SearchBot is the one tied to search features. They are separate rules |
| You need an AI file or special markup for Google’s AI answers | Documented false | “There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary” |
| llms.txt gets you into AI answers | Guesswork | A 2024 proposal. No major provider’s documentation mentions reading yours, and Google says you don’t need “new machine readable files, AI text files, or markup” |
| Schema markup makes assistants recommend you | Overstated | Google says structured data tells it about hours and departments, and that it “does not guarantee that features that consume structured data will show up” |
| ChatGPT’s local results come from Bing | Not disclosed | OpenAI’s published crawler documentation names no search or maps partner. Treat any specific claim here as unverified |
| Some percentage of search will be AI by a given year | Guesswork | Nobody who runs one of these systems publishes that number. Plan your week around it at your own risk |
Two things follow. The only documented lever on the ChatGPT side is whether you let the crawlers in. And Google has written down in plain language that there is no AI-specific optimization: to be quoted, “a page must be indexed and eligible to be shown in Google Search with a snippet”.
The crawlers you can actually control
This is the one place where a bad setting genuinely removes you from the picture. A security plugin, a firewall rule or a copied robots.txt file can block a search crawler while leaving your site perfectly visible to humans, and nothing on your site will look wrong.
| Crawler | What it does | If you block it |
|---|---|---|
| OAI-SearchBot | Surfaces sites in ChatGPT’s search features | You opt out of ChatGPT search results |
| GPTBot | Crawls content that may be used in model training | Your content is not used for training. Search is a separate rule |
| ChatGPT-User | Visits a page because a user asked something now | Robots rules may not apply; the fetch is user-initiated |
| PerplexityBot | Surfaces and links sites in Perplexity results, “not used to crawl content for AI foundation models” | You opt out of Perplexity’s results |
| Google-Extended | Gemini training and grounding only | Nothing in Search changes: it “does not impact a site’s inclusion in Google Search nor is it used as a ranking signal” |
That last row is the one most often gotten backwards: blocking Google-Extended is a decision about training, not about visibility.
Check the file, don’t assume. Open yoursite.com/robots.txt in a browser right now. If you see a blanket User-agent: * / Disallow: /, or a stack of AI bot names someone added for you, you have opted out of the thing you are trying to get into. OpenAI notes it can take about 24 hours after a robots.txt change for its systems to adjust, so fix it and check again tomorrow.
What an assistant can find out about you
Once the crawlers are allowed in, the question is what they find. Assistants describe businesses using text. Not your logo, not the photo of your van, not the PDF price list, not the hours baked into an image. Text. So the work is the same work that makes a site useful to a person in a hurry:
- Say who, what and where in words. Your site should state the business name, what you do, the towns and neighborhoods you cover, your hours, your phone number and your license number as readable text on the page.
- Put prices, process and answers in text. What a visit costs, how a quote works, what happens on the day, what you will not do. If you have an FAQ, it belongs in HTML, not in an image or a downloadable file.
- Keep the facts identical everywhere. Google Business Profile, Bing Places, Apple’s listing tool (now part of Apple Business, which “combines the capabilities previously found in Apple Business Manager, Apple Business Essentials and Apple Business Connect”), Yelp and your trade’s directories. Same name, same address format, same hours, same phone.
- Get reviews that mention services and places. Ask, don’t script. Google allows merchants to “encourage the posting of content that does represent a genuine experience, without offering incentives”, and prohibits paid or incentivized reviews, pressuring customers on the premises, and selectively soliciting the positive ones while discouraging the negative ones.
- Be mentioned by local sites. The chamber, the neighborhood association, the supplier that lists its installers, the school you sponsor, the local paper. These pages are readable by every crawler on the list above.
None of that is new advice. That is the point: the systems that summarize the web are reading the same web, and they are reading it as text.
What “specific” looks like for three businesses
Vague pages produce vague answers. Compare what an assistant has to work with:
An HVAC company. Vague: “Quality heating and cooling solutions for residential and commercial customers.” Specific: “We replace and repair residential gas furnaces, heat pumps and mini-splits in Oak Park, Berwyn and Forest Park. Diagnostic visits run until 8 p.m. on weekdays, and the diagnostic fee is credited against the repair.” The second version contains four things a customer might ask about by name.
A dentist. Vague: “Comprehensive dental care for the whole family.” Specific: “We are a general dental practice taking new patients, including children from age one. We are in-network with the plans listed on our insurance page, we hold evening appointments on Thursdays, and emergency slots are held each morning for existing patients.” Note what is missing: no promises about outcomes, no claim to be the best. Dental advertising rules vary by state board, and the specifics that help here are facts about access and scope, not superlatives.
A dog groomer. Vague: “Pampering for your furry friend.” Specific: “Full grooms for standard poodles and doodles take about two and a half hours. We groom senior dogs and anxious dogs without sedation, we do nail trims as a walk-in on Tuesdays, and we are a five-minute drive from the Riverside dog park.”
The pattern is the same in all three: name the exact service, the exact area, the timing and the constraint. Those sentences are quotable. “Quality solutions” is not. Which of them deserve their own page is a separate question, answered by the emergency, planned and required method.
What a blog does here, and what it can’t do
An honest account of blogging in this context has two halves.
What it does: it produces the specific, locally grounded text that everything above depends on. A post about what a heat pump costs to run through a Midwest winter, written with your own numbers and your own service area, is a page that can be indexed, quoted and cited. Google’s requirement for being shown as a supporting link is exactly that a page be “indexed and eligible to be shown in Google Search with a snippet”. Pages that say something specific clear that bar. Pages that say nothing in particular do not, however often they are published.
What it cannot do: change where you are, or how many people have reviewed you. Google names three factors for local ranking — relevance, distance and prominence — and says there is “no way to request or pay for a better local ranking on Google”. This part is our inference, not something OpenAI has documented: an assistant answering “who should I call near me” has to get its local businesses from somewhere, and if that somewhere is built like Google’s local results, it inherits the same limits. If a closer business with more reviews fits the request better, expect the assistant to name it, however much you blog. Content competes on relevance, not on distance.
We build the readable part. LocalBiz builds the site, then generates service pages, location pages and blog posts from the details of your actual business: your services, your area, your prices and process. It helps with the Google Business Profile work and tracks where you rank for the searches that matter. What it does not do, and what nobody can sell you, is a place in ChatGPT’s answers. It makes sure there is something specific and crawlable to find when an assistant goes looking. See what it costs.
The ten-step checklist
- Open your robots.txt. Confirm nothing blocks OAI-SearchBot, PerplexityBot or Googlebot, and that no plugin has added a blanket disallow.
- Put your business name, service area, hours, phone and license number in text on every page, usually in the footer.
- Write one page per service you would quote separately, each naming the job the way a customer says it.
- Publish your prices or your price ranges, or at minimum how pricing works. In text, on a page, not in a PDF.
- Answer your ten most common customer questions on the site, in full sentences.
- Claim and complete your Google Business Profile, Bing Places and Apple listing, plus the two or three directories that matter in your trade.
- Make the facts identical across all of them, down to the suite number and the hours.
- Ask every customer for an honest review, without incentives and without filtering who you ask.
- Get listed and mentioned on local sites: the chamber, suppliers, associations, the local paper.
- Publish something specific about your work and your area on a regular schedule, and keep a record of what you published when.
Read your homepage out loud. If a stranger could not tell from the words alone what you do, where you do it and what it costs, no software reading the page can tell either. Fixing that one sentence is a twenty-minute job, and it feeds every system on this page.
How to check whether any of this is working
There is no console for this, so you check it from both ends.
The demand side. Once a month, from a signed-out session, ask each assistant the five questions your customers actually ask. “Who repairs furnaces near [your town] on a weekend?” “What does a full groom cost for a doodle in [your town]?” Write down which businesses come back, which sources are cited, and whether anything about you is wrong. Wrong facts are the most actionable result you can get, because they usually trace back to a listing you can correct.
The supply side. Check robots.txt as above. Then search your server logs or your host’s analytics for the user-agent strings: OAI-SearchBot, GPTBot, ChatGPT-User, PerplexityBot. Seeing them means the door is open. Not seeing them after a month of publishing means something is blocking them, and that something is usually a firewall or a security plugin rather than your robots.txt.
The ordinary side. Most local demand still arrives through Search and Maps, and the same text feeds both. If you are not visible there at all, start with the Maps diagnostic, and if you are weighing a site against a profile alone, read that comparison first. LocalBiz customers get the service pages, location pages and scheduled posts generated from their own business details, which is the supply side of this list handled for them. The rest of the archive is at the blog.
Frequently asked questions
How do I register my business with ChatGPT?
You can’t. OpenAI publishes no business directory, no submission form and no listing to claim. What it does publish is a list of its crawlers. OAI-SearchBot is the one “used to surface websites in search results in ChatGPT’s search features”, and you allow or disallow it in your robots.txt file. Everything else an assistant knows about you it reads from your site, your listings, your reviews and pages that mention you.
Do I need an llms.txt file to show up in AI answers?
No provider documents reading yours. llms.txt is a proposal published in September 2024 by Jeremy Howard, and plenty of sites now publish one, but OpenAI’s crawler documentation, Perplexity’s and Google’s do not mention it. Google states the opposite directly: “You don’t need to create new machine readable files, AI text files, or markup to appear in these features.” Publishing one is harmless. Paying someone for one is not a plan.
Does schema markup make AI recommend my business?
Not by itself, and nobody documents that it does. Google says Local Business structured data lets you “tell Google about business hours, different departments within a business, reviews” and more, and warns that it “does not guarantee that features that consume structured data will show up in search results”. Its AI features page adds that there is no special schema you need for AI Overviews. Add it for accuracy, not for a promise.
Will blogging get my business into ChatGPT's answers?
It can only help with one half of the problem. Regular, specific posts give an answer engine quotable text about what you do, where you work and what it costs, which is what it needs to describe you at all. They do not move the other half. Nobody outside OpenAI can show you how an assistant picks a local business, but if it leans on local results the way Google builds them, a closer business with more reviews that fits the request better is the one it describes.
Sources
- OpenAI — Overview of OpenAI crawlers (OAI-SearchBot, GPTBot, ChatGPT-User)
- Google Search Central — AI features and your website
- Google Search Central — Google crawlers (Google-Extended, GoogleOther)
- Google Search Central — Local Business (LocalBusiness) structured data
- Google Business Profile Help — Improve your local ranking on Google
- Google Maps user contributed content policy — fake engagement
- Perplexity — PerplexityBot and Perplexity-User
- llms.txt — the proposal site
- Apple — Apple Business guide (Apple Business Connect)