Every business has a competitor list it has agreed on internally. It is on a slide somewhere, it has three or four names on it, and it is usually a list of the companies that turn up in deals.
Models have their own list, and nobody consulted you about it.
By the time a model is comparing options, it has already decided what kind of thing you are. That decision comes from how you are written about across the web — reviews, forum threads, listicles, your own copy — and it is stickier than any individual answer.
If the grouping puts you next to the budget options, every subsequent answer inherits that framing. You will be compared on price, because that is what the set you are in gets compared on.
The interesting output of any competitor analysis is not confirmation that your known rivals appear. It is the name you have never heard of that keeps turning up beside yours.
Sometimes that is a smaller company punching above its weight in exactly the questions you care about. Sometimes it is a company you would not consider a competitor at all, because it solves the problem a different way — and the model has decided customers might reasonably pick either. That second case is worth more attention than it usually gets, because it tells you what job customers think you do.
If a rival you take seriously never appears in answers where you do, they are not competing for that question. They may be winning elsewhere, or they may have no presence in this channel at all.
Either way it is a real asymmetry, and it is the sort of thing that is invisible from a rankings report, where you would both simply appear as blue links a few positions apart.
Group membership is not fixed, but it does not respond to a change in your homepage headline. It responds to accumulated evidence: the comparisons that exist, the categories you are listed under, the language other people use about you.
Moving from one set to another means giving the model something else to read, in enough places that the new framing outweighs the old one. That is a year of work, not a quarter. The value of tracking it is knowing whether the work is landing while it is still cheap to change course.
A competitor set read once is a snapshot, and a mildly interesting one. Read repeatedly, it becomes the most sensitive early indicator available in this channel.
New names entering the set is the signal to watch. A company that appears beside you in one question, then three, then across a whole category of questions, is establishing itself in the material that models read — and it is doing so well before it shows up as pressure in your pipeline.
Names leaving are worth noting too, though they are easier to misread. A competitor fading from answers may be losing ground, or may simply have stopped being written about while continuing to trade perfectly well. Absence from the conversation and absence from the market are different things, and only one of them is visible here.
You will not have one competitor set, you will have several, and they will not agree.
The set that appears in awareness questions tends to be broad and to include category-adjacent options. The set in comparison questions is tighter and usually closer to your internal list. The set in purchase-intent questions is often tighter still, and occasionally contains a surprise — a name that only surfaces when someone is ready to buy.
Averaging these together produces a competitor list that describes none of the situations you actually face. Keeping them separate is more work to read and considerably more useful, because it tells you where in the journey you are being displaced.
The temptation on seeing an unfamiliar name is to respond to it directly — a comparison page, a battlecard, a paragraph on the pricing page.
That is usually premature. A name appearing in a handful of answers over a week may not appear at all the following month. Wait for it to persist across different phrasings and different questions before you build anything around it, and in the meantime treat it as what it is: a hypothesis about who customers think you are like.
Contents
The channel where your brand is decided but no traffic arrives, and why the usual analytics can't see it.
Read →AI answers are replacing search clicks, and the tools to measure it were priced for enterprises. So we built one for the rest of us.
Read →Not your homepage. A working theory of the sources that shape what AI answers say about a brand, and what that means for where you spend effort.
Read →Tell us your domain and we'll send you a visibility report — free, no card.