Version française →

ChatGPT Doesn't Know My Business, Even When I Search My Own Name

You open ChatGPT, type your company's name out of curiosity, and the answer comes back vague, generic, or flat-out wrong. Yet your business exists, it has a website, customers, sometimes years of history. Here's why an AI can completely ignore it, and what that actually means for you.

It's an easy test, and plenty of business owners try it one evening out of curiosity, not expecting it to go badly. They type their company's name into ChatGPT, expecting something roughly like a Google listing. Instead: nothing much, or a vague "I don't have specific information about this business," sometimes even confusion with a similarly named company 300 miles away. The instinct is to blame the AI for being unreliable. The real issue lies elsewhere.

What this actually signals

If ChatGPT can't say anything precise about your business, that's not an isolated glitch — it's a signal about your online presence as a whole. A generative AI doesn't "visit" your business like a customer would. It was trained on massive amounts of text, and for recent or specific questions, it also searches the live web at the moment you ask. Either way, it can only reproduce what already exists, somewhere, in a readable and consistent form.

In practice, if your business only shows up on a handful of thin pages, if what you actually do is never described in plain language anywhere, or if the information floating around about you is contradictory from one source to the next, the AI simply has no reliable material to draw from. It won't invent a profile to be nice — it'll stay vague, or fall back on whatever it can find, sometimes a competitor with a similar name.

Why this happens even to solid businesses

This is often where the confusion runs deepest: businesses that are doing well, with loyal customers and a strong local reputation, discover they're nearly invisible to AI. The reason is almost always the same. Their reputation was built through word of mouth, referrals, on-the-ground trust — channels Google eventually learned to capture indirectly over the years (reviews, local citations, mentions), but that generative AIs don't read the same way.

An AI needs text, somewhere, that clearly states who you are, what you do, for whom, where, and why someone should choose you. A three-page site that hasn't been updated in years, a vague "About" page, or a business that's only ever described in person and on the ground — that's exactly the kind of black hole where an AI finds nothing to cite. This is where the technical term behind this work comes in, once the cause is clear: GEO, or Generative Engine Optimization — the practice of making a business readable and citable by generative AIs, much like SEO has long done for Google.

What this actually changes for you

The issue goes well beyond curiosity about your own name. More and more people who used to search for a provider on Google now ask an AI directly: "who does this near me," "which company would you recommend for this kind of project." If the AI doesn't know you when asked about your name specifically, chances are it won't suggest you spontaneously either when the question is broader. You're not just losing an answer to a curious query — you're losing a spot in a recommendation that, increasingly, replaces the first instinct of "let me Google it."

What you can check yourself in 5 minutes

1. Open a new conversation in ChatGPT and ask: "What do you know about [your business name]?" Note whether the answer is precise, vague, missing, or wrong.
2. Then ask a question without naming your business: "who would you recommend for [your industry] in [your city]?" See if you appear, at what position, and who shows up instead of you.
3. Repeat both questions with Perplexity and Gemini: results can differ across AIs depending on their sources.
4. Finally, search for your business the way an AI would: beyond your own website, how many third-party pages clearly describe what you do, where you operate, and what sets you apart?

If the answers are thin or wrong, that's not a dead end — it's a lack of usable information, and it can be fixed. That's exactly what we look at in an audit.