Why Is My Brand Not in ChatGPT?

One symptom, five causes, no error message

You ask a question your business should own, your name does not appear, and nothing explains why. That silence has five separate causes and they are indistinguishable from where you are standing, which is the entire difficulty.

This is unlike almost every other channel. A page that does not rank still leaves evidence in a search console. An advert that does not convert leaves a click and a bounce. Here the whole event happens somewhere you cannot observe, and the output is a paragraph that simply mentions somebody else.

So the work is diagnostic rather than corrective. Before changing anything, establish which of five things is true, because four of them are cheap and one is a quarter of work, and picking wrong costs you the quarter.

The order below is by cost to test, not by likelihood. Each check is cheaper than the one after it, and a failure at any stage makes every later stage unmeasurable. Testing them in this order is the only sequence that does not waste effort.

The five reasons, in order

Take them in sequence and stop at the first one that is true for you.

1. Nothing can reach your pages

A rule somewhere refuses the crawler that feeds the assistant. This is binary, it produces complete absence, and it is set in files you control. It is also the most common cause of a business being entirely missing rather than merely under-represented.

The reason it survives for years is that nobody chose it. The rule arrives with a security product, a bot-management default or a hosting setting, written by somebody solving a different problem, and no notification is ever sent. Reading the rules directly is the only way to know.

2. You are reached and passed over

The pages are fetched and contain nothing that can be lifted. Commercial writing sets context, acknowledges nuance, and states the point in the middle. A system assembling a short answer needs the point near the top and needs it in one piece.

This produces partial absence rather than total. You appear occasionally, usually for your own name, and never for the questions that matter.

3. There is not enough of you on the subject

One page about a topic is a page. Reaching for you as the source on a subject is a judgment about whether you are a reasonable thing to reach for, and a single page is thin evidence.

4. Nobody outside your site says it

Assistants draw heavily on material that is not yours. If your category is discussed and you are absent from the discussion, you can do everything right on your own domain and still not be named.

5. You are asking the wrong question

This is the one that costs nothing and is skipped most. Businesses test using their own vocabulary. Buyers describe a problem. If you ask about your category name and a buyer would ask about their symptom, you have measured the wrong thing.

How to tell which one is yours

Each cause has a distinguishing test that takes minutes rather than weeks, and running all five takes an afternoon.

For access, read the rules on your site and at the edge, and list which named clients are permitted. Do not ask a colleague what the setting is. The whole point of the check is that the configuration and the intention have come apart.

For answerability, read only the first two sentences of your ten most important pages. If they do not answer the question in the title, this is at least part of your problem. This test is uncomfortable because the pages that fail hardest are usually the ones somebody spent the most on.

For depth, count the pages that answer a distinct question within your priority subject. Pages that mention the subject do not count. If the honest number is one or two, depth is a real constraint and it is the slow one.

For outside presence, search the way a buyer describes the problem and read what comes back. If the discussion exists and you are not in it, you have found something no amount of on-site work resolves.

For the question itself, write down how a customer described their problem to you the last time one did, in their words, and ask that. Businesses are consistently surprised here, because the phrasing a company uses about itself is rarely the phrasing anybody uses to look for it.

The reason nobody checks first

Reason five deserves its own section because it is free to test, it is frequently the answer, and almost nobody starts there.

A business asks an assistant something like "what are the best providers of" followed by the category name it uses internally. The category name is often a term the industry adopted and customers never did. The answer comes back naming other companies, and the conclusion drawn is that the business is invisible.

What has actually happened is that a question nobody asks has been answered by whoever wrote about that phrasing, which tends to be other suppliers rather than anybody a buyer would find useful. The result says almost nothing about your visibility to actual buyers.

The correction is to build the question set from language you did not write. Support tickets, sales call notes, the subject lines of inbound emails, the phrasing in a review. Any of those will produce questions that differ from your marketing vocabulary, often startlingly.

Then ask those. It is common for a business that appeared absent to be named regularly once the questions are phrased the way a person would phrase them, and equally common for the reverse to be true. Either result is worth more than the one you get from asking about yourself in your own words.

What each fix actually costs

The five causes differ enormously in what it takes to resolve them, and knowing that in advance stops a cheap problem being treated as an expensive one.

Access is minutes of work by somebody with the right credentials, followed by waiting for pages to be refetched. There is no design involvement and no approval cycle. If this is your cause, you are days from a different situation.

Answerability is days of writing and weeks of approval. The writing is genuinely fast because you are relocating conclusions you already reached. The approval is slow because the openings in question were usually argued over once already.

Depth is a quarter of sustained attention on one subject, and it resists compression. Adding people does not shorten it much, because the pages have to be genuinely different from one another and that needs somebody holding the whole subject in mind.

Outside presence has no schedule at all. It is a consequence of doing things worth writing about and being easy to reach when somebody wants to write about them. Any supplier quoting a timeline for it has misunderstood what it is.

The question set costs nothing and takes an hour. It is the cheapest item on the list by an enormous margin and it is the one most often skipped, which is a fair summary of how this work usually goes wrong.

Three things not to do about it

Some responses to this problem are actively counterproductive, and they are common enough to be worth naming.

Do not write a page about your own brand name. A page explaining who you are answers a question only asked by people who already know you exist. It will not help you be named in an answer to a question about the category, and it consumes the effort that would have gone somewhere useful.

Do not commission volume. The instinct on hearing that depth matters is to order more pieces. Forty pages across forty subjects reads as a large site and behaves as a thin one, because retrieval happens per subject. One subject covered properly beats ten covered adequately, and that is not the answer a content calendar wants.

Do not treat one reading as a result. Asking once, seeing a rival named, and reorganizing a quarter around it is building on a single sample of something that varies. Ask repeatedly across days before concluding anything, and record what you got each time.

The pattern behind all three is the same. The absence feels urgent, and urgency produces action on whichever cause is most convenient to act on rather than the one that is true.

When the answer is that nothing is wrong

There is a sixth possibility, and it is worth stating plainly because it is the one nobody sells a service for.

Some categories are heavily entrenched. The sources that get drawn on for a subject may be long-established publications, and the reason you are not named may be that you are newer than everything being cited, with no defect involved anywhere.

The honest read on this is to look at who is being named rather than at whether you are. If the answers consistently reach for organizations decades older than yours with orders of magnitude more written about them, then your absence is a fact about the category rather than a fault in your site.

That is not a reason to do nothing. It is a reason to pick a narrower subject where the entrenchment does not exist, which is a strategic decision rather than a technical one. Being the obvious source on a specific question is achievable in a way that being named alongside the largest names in a broad category is not.

Anybody who tells you every category is equally winnable has not looked at who currently gets cited in yours.

Why the order is not negotiable

It is tempting to run all five checks at once and act on whatever looks worst. That produces the wrong answer often enough to be worth arguing against directly.

A failure at any stage makes every later stage unmeasurable rather than merely less important. If the crawler is refused, then your pages are not being passed over for being unquotable, because they are not being read at all. Rewriting them teaches you nothing, and when the access rule is eventually corrected you will have no idea which of the two changes produced the result.

The same applies further down. Judging your depth while your content sits behind an interface that hides it from anything not running a browser is judging a site nobody has seen. The measurement is not slightly off, it is measuring something else entirely.

There is also a budget argument. The stages are ordered by cost, so working in order means you spend nothing until you have ruled out the free explanations. Working out of order means the expensive explanation gets funded first, because it is the one with a supplier attached to it and a plan that looks like progress.

Our page on the AI visibility checklist sets the same sequence out as fourteen concrete checks if you want the version you can work through directly.

What actually changes first when you fix it

Knowing what recovery looks like helps you tell a real improvement from noise, and the shape is fairly consistent.

The first thing to move is usually your own name. Questions that mention you directly start returning accurate descriptions rather than nothing or something garbled. This is a low bar and it is genuinely informative, because it tells you retrieval is now working even if selection is not.

The second thing to move is long-tail questions with specific wording, particularly ones that match a page you rewrote. These are the easiest to win because fewer sources address them precisely, and they are the earliest evidence that a rewritten opening did something.

The last thing to move, if it moves at all, is the broad category question with the most obvious phrasing. That one is contested by everybody and usually answered from long-established sources. Treating it as the measure of success means concluding that nothing worked for several quarters while a great deal is working.

Watch all three separately rather than as one percentage. A number that averages a question you can win with a question you cannot will move slowly and tell you very little about which of your changes mattered. There is more on this in our piece on the AI visibility score.

Bottom line

Check in order and stop at the first thing that is true. Access takes minutes and explains most complete absences. The question set takes an hour and explains a surprising number of the rest.

Do not start with depth. It is the most expensive cause, it is the one every supplier is equipped to sell you, and committing a quarter to it while a firewall rule refuses every request is the most costly mistake available here.

And ask more than once before you conclude anything at all. A single reading of something that varies is not a measurement, and every decision built on one inherits that.