ChatGPT doesn't know my business exists. Here's exactly why.

Six specific reasons an AI assistant will miss a real, good business when a customer asks who to hire — and why the reason that applies to you depends on your industry.

Jarvis Report · 2026-08-03 · built from live public data — we never invent a score

Ask ChatGPT for the best plumber in Phoenix and it answers with 3 names, stated as fact. Not a results page, not a map with twenty pins — 3 names. Every business missing from that answer just lost a customer who will never know it existed. Owners who search chatgpt doesn't know my business after watching that happen usually assume one big cause. In our audits it is almost never one cause — it is one of six, and which one depends on the industry.

An AI answer is a shortlist, not a search page

The old failure mode was ranking on page two of Google. Painful, but survivable — the page still existed, and a motivated customer could find you. The new failure mode is different in kind. When a customer asks an assistant who to call, the assistant composes a closed answer: a few names with a sentence of justification each. There is no scroll, no page two, no "more results." You are either in the answer or you are not a fact about the world.

That shortlist is also unstable. Ask the same buyer question twice and the names can change between runs, which is why any serious measurement asks the question repeatedly instead of trusting a single answer. A single lucky mention is not visibility; a single absence is not proof of invisibility. The pattern across repeated runs is the real signal — and one more distinction matters: an assistant that declines to recommend anyone is a different result from an assistant that recommends your rivals and skips you. Collapsing those two states into "AI ignores me" hides what is actually wrong.

Two different ways an AI can know you

Assistants draw on two supplies of information, and a business can be missing from either or both.

Model memory. What the underlying model absorbed from the public web during training. This is slow-moving and biased toward businesses that have existed for years, been written about, listed, reviewed, and linked. A newer business, or one that lives mostly on social platforms, can be simply absent from it.

Live retrieval. What the assistant looks up on the web at the moment of the question. This is where most local recommendations actually come from — and it means the assistant is only as informed as the pages its crawler can fetch and read right now. Which brings us to the six reasons.

The six reasons AI skips a real business

1. Your website is invisible to the bots that feed AI

The crawlers behind AI systems generally do not execute JavaScript. A site built as a modern app can look complete in a browser and arrive at the crawler as a nearly empty shell. We know because we measured our own site doing it: every page served the same 5,034-byte empty frame to AI crawlers while humans saw the full product. To those bots, our pages said nothing at all — and a bot that receives nothing has nothing to recommend. If your site was built in the last few years on a modern framework, this is the first thing worth ruling out, because no amount of review-building fixes a page the reader cannot open.

2. Your identity doesn't resolve to one business

Engines assemble a picture of a business by merging mentions from many public surfaces. That merge runs on your name, address, and phone number matching. An old address on one directory, a tracking phone number on another, "& Sons" spelled three ways — each mismatch splits you into fragments that look like separate, weaker businesses. None of the fragments has your full history, so none of them earns a place in a shortlist. This failure is quiet: every individual listing looks fine on its own, and the damage only appears when you compare them against each other.

3. Your public proof is too thin to make a shortlist

When an assistant must pick a few names out of dozens, it leans on legible, comparable evidence — review count, rating, and how recently people have reviewed you. This is unfair in the way all proxies are unfair: a 4.9-star business with 40 reviews reads as less proven than a 4.4-star business with 4,000. The assistant is not judging your work. It is judging the paper trail your work left in public, and a thin trail loses to a thick one almost every time.

4. Nothing about you is machine-readable

Pages written for humans still need a layer that machines can parse without guessing — what you do, where you are, what hours you keep, what you charge. Businesses that state these facts in a form software can read remove every excuse an engine has to be unsure about them. Businesses that leave it to inference get described vaguely or mis-categorized, and a business the engine cannot confidently categorize is a risky thing to put in an answer. Engines prefer to skip a maybe.

5. You're missing from the surfaces AI reads for your industry

Assistants do not consult the same sources for a pizzeria as for a probate lawyer. They pull from wherever the buying evidence for that category lives — and if you are absent from the two or three surfaces that dominate your category, being present everywhere else buys you little. The table below shows how differently this plays out. It is also why generic advice fails: the fix list for a restaurant is nearly useless to a roofer.

6. You've never been the answer anywhere else

Engines learn who the answer is partly from who has already been the answer: local press, "best of" round-ups, community forum threads, suppliers' partner pages, chamber listings. These citations are how a name stops being a listing and starts being a reputation. A business that has never been mentioned by anyone else is asking the assistant to vouch for it first. Assistants are built to be cowards about exactly that.

Different industries, different evidence

The six reasons are universal; their weights are not. What follows is the pattern we see in audit data across categories — where the answers for each industry are actually assembled from.

Read the middle column for your industry and an uncomfortable question follows: do you actually know what those surfaces say about you today? Most owners do not, because each surface looks fine in isolation. The problems live in the gaps between them — and in what the crawlers receive, which no human ever sees by browsing.

So which of the six is hiding your business?

This is the part that resists a generic article, and we will not pretend otherwise. The six causes look similar from the outside but they are diagnosed differently: one requires seeing your site the way a crawler sees it, another requires comparing your identity data across surfaces character by character, another requires putting your review evidence next to the rivals AI actually names in your city, and the industry weighting changes which check comes first. Owners who guess tend to fix the visible thing — usually the website — while the thing actually hiding them sits in a directory they have not opened in years.

Diagnosing it is a measurement job, so we built the measurement: Jarvis Report runs the real buyer questions for your city, reads your listing the way the engines do, pulls your rivals' live numbers, and shows which of the six reasons applies to you — from live public data, in about a minute, free. If you want to see how deep the full report goes first, the pricing page shows what stays free and what the dashboard adds.

Live public data only. We never invent a score.

Questions owners ask us

Why does ChatGPT recommend my competitor and not me?

Because on the evidence the assistant can read — identity consistency, review mass, machine-readable pages, presence on your category's key surfaces — your competitor is currently stronger or clearer than you are. That is a statement about public data, not about the quality of your work. It is also fixable, which is precisely why it is worth diagnosing correctly.

Is this the same problem as ranking low on Google?

No, and the difference is the whole point. A low rank leaves you findable by anyone who keeps scrolling. An AI answer has no scroll. When owners tell us "chatgpt doesn't know my business," what they have usually discovered is this harder version of invisibility: absence from a closed answer that customers treat as complete.

My website looks great. How can it be the problem?

Because "looks great" is a claim about what a browser renders, and crawlers are not browsers. If the content arrives only after scripts run, most AI crawlers never see it. The only way to know is to check what the bot receives — which is one of the first things our scan does, because we were burned by exactly this on our own site.

Can I just pay someone to fix all six at once?

You can pay for a lot of motion, but without a diagnosis you are betting money on the wrong cause in an unknown number of cases. The causes are cheap to distinguish once measured, and the measurement is free. Diagnose first, then spend where your industry's evidence actually lives.

How often should I re-check?

AI answers move: models update, rivals accumulate reviews, directories go stale on their own. Treat visibility like a number that drifts, not a box you tick once. If you typed "chatgpt doesn't know my business" once, put a repeat check on your calendar — the answer you fix this month can quietly change by next quarter.