There is no submission form that guarantees your business will appear in ChatGPT Search. The practical job is to make your website publicly accessible, allow OpenAI's search crawler where appropriate, publish useful information that answers real questions and build enough credibility that your pages are worth surfacing as sources.
This is already becoming a familiar question.
A business owner asks:
How do I rank in ChatGPT?
The temptation is to invent a completely new branch of marketing around that sentence.
I would resist that.
There are some genuinely new technical details.
But a lot of the work looks remarkably like good web publishing and good SEO.
ChatGPT Search can cite websites
OpenAI's current Search documentation says ChatGPT can search the web and include links to sources.
That means a business website can potentially appear as a cited source inside an answer rather than only as a traditional blue search result.
OpenAI also says any public website can appear in ChatGPT Search.
That is important.
There is no requirement to be a large publisher or an official OpenAI partner.
There is also no guarantee.
Do not block OAI-SearchBot if you want to be discoverable
This is one of the concrete technical steps.
OpenAI says website owners who want their content to be discovered, surfaced and clearly cited should make sure OAI-SearchBot is not blocked.
That means checking:
robots.txt;
firewall rules;
CDN/bot-protection rules;
any security layer that may return errors to automated crawlers.
A crawler that cannot access the page cannot properly read it. That is why the basic work of getting a business website discovered and indexed still matters.
This is especially relevant for sites using aggressive bot protection.
A technically public page can still be effectively invisible to a crawler if the infrastructure blocks the request.
OAI-SearchBot and AI training are not the same question
This distinction is worth understanding.
A website owner may want public pages available to search without making every possible AI-use decision identical.
OpenAI documents separate controls for different crawler purposes.
The important point for this article is narrow:
if you want the site available for ChatGPT Search discovery and citations, do not accidentally block the search crawler you expect to read it.
Do not copy a random robots.txt file from another site without understanding what it blocks.
Crawlability only gets you through the door
Allowing a crawler does not make a page useful.
You can have perfect technical access and still publish weak content.
A page saying:
We provide innovative digital solutions tailored to your needs.
does not give an answer engine much to work with.
A page clearly explaining:
what the business does;
who it serves;
how pricing works;
where it operates;
what is included;
what common customer questions mean;
creates much more usable information.
The same principle already matters in ordinary search.
AI search simply makes the source-quality question more obvious.
Answer actual questions
If customers repeatedly ask:
How much does this cost?
Do I need hosting?
Can I move my website?
Do you work with businesses in Longford?
Should I use WordPress?
those questions can become useful content.
That is exactly why the Built by Gavin Guides library exists.
The goal is not to publish hundreds of pages because AI exists.
It is to create a body of information that is clear enough to be useful to:
a customer;
Google;
ChatGPT Search;
and any other system trying to understand the subject.
Clear answers near the top help humans first
I like putting a direct answer near the beginning of a guide.
Not because there is a secret ChatGPT formula demanding it.
Because it is good communication.
A visitor searching a specific question should get the answer quickly.
The deeper explanation can follow.
That structure also makes the page easier for search systems to interpret because the main point is not buried beneath six paragraphs of introduction.
Original judgement matters
AI systems have access to enormous amounts of generic material.
That makes commodity content less interesting, not more.
If every web designer publishes:
Five reasons your business needs a website
and each article says essentially the same thing, there is little reason to remember any particular source.
A more useful page contains decisions.
What would I recommend?
When would I not recommend it?
Where do people overpay?
Which trade-off matters?
What has been learned from actual small-business work?
That is one reason I push so hard against generic website copy.
A genuine point of view makes the information more distinctive.
Build a connected site, not isolated articles
Internal links matter because they help people and crawlers understand relationships.
A guide about website hosting can link to:
static websites;
Cloudflare;
monthly fees;
maintenance;
ownership.
That creates a navigable subject structure.
The article is no longer an isolated page floating in the site.
It belongs to a useful body of knowledge.
This is normal information architecture.
It also makes sense for AI-era discovery.
Your business pages still matter
Do not make the mistake of creating a brilliant knowledge base while the actual commercial site remains vague.
An AI answer may surface an article.
The person may click it.
Then what?
They still need to understand:
who you are;
what you sell;
where you work;
what it costs;
and how to contact you.
AI visibility without a credible destination is not much of a business strategy.
Can structured data help?
Structured data can help machines understand page types and entities where supported.
It is sensible web implementation.
But there is no magic "ChatGPT schema" that guarantees citations.
Use appropriate markup because it improves the clarity of the site.
Do not buy an AI SEO package based on a claim that one hidden JSON block will suddenly make ChatGPT recommend you.
Do I need an llms.txt file?
That question deserves its own guide later in this cluster.
The short answer is that there is no general reason to treat llms.txt as the main path into ChatGPT Search.
For OpenAI Search, the current documented crawler-access requirement is about OAI-SearchBot.
Start with the thing the platform itself actually documents.
Can I submit my website directly?
There is currently no equivalent of a guaranteed "submit this page to ChatGPT and get indexed" button for ordinary publishers.
OpenAI says public websites can appear and gives crawler-access guidance.
That is different from promising inclusion.
If somebody sells guaranteed ChatGPT citations, I would ask exactly what control they believe they possess.
Gavin's take
I would not build a separate “ChatGPT strategy” that ignores the website fundamentals. If a page is hard to crawl, vague about the business or adds nothing beyond generic copy, a new AI-search label does not rescue it. I would first make the site genuinely useful and technically accessible, then make sure the relevant search crawlers are not being blocked.
How would I approach this for a small business?
I would work in this order.
First, make sure the site is technically accessible.
Then make sure the core commercial information is strong.
Then build useful content around real customer questions.
Then strengthen the site through internal links, proof and real-world authority.
Then measure referral/search behaviour where possible.
That is less exciting than declaring SEO dead.
It is also much more defensible.
What I'd do
For a small business, I would prioritise the same assets that help across search: clear service and location pages, useful first-hand Guides, sensible internal links, strong business information and a crawlable site. Then I would check the specific crawler controls for services such as ChatGPT Search rather than chasing speculative “AI optimisation” tricks.
Gavin’s perspective
Built by Gavin's approach
The Guides strategy is deliberately built for both traditional and AI-powered discovery.
But I am not creating separate "ChatGPT pages" full of artificial phrasing.
The same article should be useful to the business owner reading it directly.
That is my test.
If the page only exists to manipulate a machine, it is probably weak content.
If it genuinely answers the question, explains the trade-offs and connects to deeper useful material, it has a reason to exist regardless of which search interface surfaced it.
The bottom line
To improve your chances of appearing in ChatGPT Search:
make the website public;
allow OAI-SearchBot where appropriate;
avoid infrastructure that accidentally blocks it;
publish clear, original, useful information;
build sensible internal links;
and give the business enough real credibility to be worth referencing.
There is no guaranteed ranking switch.
Anyone telling you otherwise is selling certainty OpenAI itself does not offer.
Want your website structured for modern search without chasing gimmicks?
See how I approach small-business websites or tell me what you want the site to become known for.
The best starting point is usually useful information and clean technical foundations — not an "AI visibility" add-on.
---