Field Guide
Why AI doesn’t mention your brand
You ask ChatGPT for the best tool in your space and it names three competitors and leaves you out. Frustrating, and also useful, because the reason is usually one of five. Here they are, how to tell which is yours, and what actually moves each. The first step in every case is the same: see the current answer before you try to change it.
1. The model has never really heard of you
This is the most common case, and the hardest to hear. If the web barely describes you, the model has nothing to assemble an answer from. This is where new companies and well-funded but quiet ones sit. You’ll spot it because the model either omits you entirely or says something generic and slightly wrong when you name yourself directly. You don’t fix this on your website. You fix it with presence: reviews, communities, and third-party articles that give the model something to learn you from.
2. It knows you, but not for this
The model has you filed under one category and the question is about another. A tool known for invoicing gets skipped on “expense management” even though it does both. You’ll see it when you’re named on some prompts and invisible on adjacent ones. So say the second category plainly, on your site and in your llms.txt, and earn a few third-party mentions that file you under it.
3. Your description is ambiguous
Your homepage leads with a slogan and the model can’t tell what you sell. “Reimagining how teams move” tells a language model nothing. This one is entirely yours to fix, and it’s the fastest win on the list. State your category in a buyer’s words, high on the page, and put it in a clean llms.txt.
4. Competitors own the sources the model trusts
You exist, you’re clear, but your rivals are all over G2, the “best tools for X” roundups, and the Reddit threads, and you’re not. The model repeats the consensus it was shown, and right now that consensus is them. You’ll recognise it when the same two or three competitor names come back on every prompt. Closing it is the slow work: claim your review profiles, get into the roundups that already rank, and show up helpfully where your buyers actually talk.
5. The model’s information is stale
You rebranded, pivoted, or shipped the feature that changes the answer, and the model is working from an older snapshot. This corrects over time as the web catches up and the model retrains, and you can speed it along by making the new reality loud and consistent everywhere a model looks, so the fresher signal outweighs the old one.
Telling them apart
You can’t diagnose this from one prompt in one model on one afternoon. The signal is in the pattern across models and across phrasings: named or not, filed under the right category or not, and which competitors keep appearing in your place. That pattern is exactly what a scan is for, and it turns “AI ignores us” into a specific one of the five above that you can actually work on.
Find out which one is yours
A free meraGEO scan runs your brand across ChatGPT, Claude, and Gemini and shows where you’re named, where you’re skipped, and who gets named instead, which is enough to tell which of the five you’re facing.
Run a free scan