Why AI Cites Your Competitor and Not You (and How to Close the Gap)

The sting: they are everywhere, you are nowhere

It is a specific kind of frustrating. You ask ChatGPT a question in your own category, and there is your competitor, named and recommended, sitting in the answer as though it were obvious. You try the same thing in Perplexity, reworded, and they are there again. You are not. From the buyer's side, the shortlist looks authoritative, which means your absence reads as a verdict rather than a gap.

The instinct is to assume you have done something wrong, or they have some trick. Neither is usually true. The models are pattern matchers, and they are reflecting a pattern that already exists out in the world. Your competitor is being cited because, in the data the model learned from and retrieves against, your competitor is the more established answer. That is a problem you can work on, once you know what actually feeds it.

Why it is not what you think

Most people assume the fix is more: more blog posts, more keywords, more backlinks. The data does not support that, and this is the most important thing to understand before you spend a dollar closing the gap.

When Ahrefs studied around 75,000 brands to see what correlated with visibility in ChatGPT, AI Mode and AI Overviews, the strongest signals were not the classic SEO ones. As their analysis of AI brand visibility factors found, branded web mentions and mentions on high traffic platforms tracked closely with AI visibility, while traditional metrics like backlinks, domain rating and sheer content volume showed notably weaker correlations. In other words, the thing you have been grinding on may be the thing that matters least.

That reframing is echoed in the argument that brand authority now beats topical authority in AI search. The piece makes an uncomfortable point well: authority is what others say about you, not what you publish about yourself. Piling up keyword-targeted articles built a lot of what the author calls content landfill, and models are not impressed by volume. They are impressed by demand and recognition that exists beyond your own domain.

The real reasons AI cites your competitor

Strip away the myths and the actual drivers are consistent. Your competitor tends to win on some combination of the following, and usually several at once.

They are mentioned more across the web

The clearest driver is simple presence in the wider conversation. Your competitor is written about, referenced and discussed in places you are not, and those mentions are the raw material a model draws on when it decides who to name. This is why the Ahrefs correlations pointed so strongly at branded mentions rather than on-site signals. The web is effectively voting on who the answer is, and your rival is getting more votes.

Crucially, many of these mentions are not links at all. A model can absorb your competitor's name from an article, a video transcript or a comparison piece without a single hyperlink involved, which is exactly why a link-first strategy underperforms here. What you want is to be named, in context, in the places your buyers and the models both read.

It also helps to make your own pages easy to lift a clean answer from, the discipline covered in crafting content that gets cited by LLMs, so that when a model does reach you, it can quote you without friction rather than skipping to a source that made the job easier.

They have brand authority, not just content

There is a difference between having published a lot and being known for something, and models lean on the second. A brand with genuine demand, the kind that shows up as people searching for it by name, carries a signal that no amount of self-published content replicates. That demand tells the model your competitor is a real, chosen answer in the market, not just a prolific publisher.

This is the hard part for a challenger, because you cannot manufacture recognition overnight. But you can build it deliberately, and the businesses that do tend to see their citations follow. Understanding how answer engine optimisation builds brand trust in the AI era is the starting point for turning content effort into actual authority.

They own the topic, not just keywords

Citations cluster around brands that cover a topic with real depth, not ones that spike on a single question. When Semrush analysed more than a thousand categories across 50,000 brands, it found that AI visibility is a topic-level game, and its study of topic authority in ChatGPT showed that traditional SEO strength explained who owned a topic only about half the time. Winning meant consistent, deep coverage across a whole cluster of related questions.

There is real opportunity buried in that finding, because most topics have no clear owner yet. If your competitor is being cited, they may simply have gone deeper on the topic than you have, and depth is something you can build. Our approach to entity and topical authority for AI search is built around exactly this kind of cluster-level coverage.

They are co-mentioned with the category leaders

This one is subtle and powerful. Models recommend brands they see grouped with the names that define a category, and being recognised is not the same as being recommended. Search Engine Land's analysis of what co-mentions reveal about the AI recommendation gap found that brands appearing alongside category leaders in third-party content were recommended far more often than brands the model understood equally well but never saw in that company.

The example is stark: in one category, a leading brand appeared in the majority of recommendations while direct rivals with near-identical profiles appeared in none, and the difference was co-mention density in outside coverage. The lesson for a challenger is to earn placement in the comparisons, round-ups and editorial pieces where the leaders already appear. You want the model to keep seeing your name in the same sentence as theirs.

They are a clearer entity to the model

Finally, your competitor is often simply easier for the model to understand. AI systems work by resolving a mention to a known entity, a process the Wikipedia entry on entity linking describes as connecting a name in text to a unique identity in a knowledge base. When your competitor's name, category and attributes are consistent everywhere, that resolution is clean and confident. When yours are patchy or contradictory, the model hesitates, and hesitation reads as absence.

Ambiguity is quietly expensive here. If two businesses share a similar name, or your details differ between your website, your listings and third-party profiles, the model cannot be sure which mentions belong to you, and it tends to resolve that doubt by naming the competitor it is sure about instead.

A clear entity is the foundation everything else sits on, because mentions and co-mentions only help if the model can reliably attach them to you. Understanding how ChatGPT selects and attributes its sources shows why entity clarity is the difference between a mention that counts and one that floats free.

The good news: most of your market is invisible too

Before this reads as an impossible mountain, here is the encouraging part. The field is wide open. A study covered by Search Engine Journal found that around 90% of brands had essentially zero AI search mentions, with only a small fraction earning any visibility at all across the platforms tested.

That means your competitor is not up against a saturated market, and neither are you. A handful of brands are winning the citations in most categories, which is daunting if you are behind them but hugely encouraging against everyone else. Acting now, while the majority are still invisible, is far cheaper than trying to break through once the space is crowded. The gap is real, but it is early, and early gaps are the ones worth closing.

This is the window our AI search work is built around, because a citation won while a category is still open is far harder for a latecomer to take back than one fought for in a crowded field. The brands moving now are buying themselves a lead that compounds.

How to close the citation gap

The plan follows directly from the causes. Since citations track mentions, the work is earning visibility beyond your own website. Pursue placement in the comparisons, round-ups, industry pieces and videos your buyers and the models read, and aim specifically to appear alongside the category leaders rather than in isolation. Co-mention density is something you can build with deliberate digital public relations and partnerships.

In practice that means the unglamorous work of getting named where it counts: pitching yourself into the industry round-ups and comparison pieces buyers already read, earning mentions in podcasts and videos, contributing expert commentary to publications in your space, and building the partnerships that put your name next to the leaders. None of it is a quick hack, but all of it compounds, because every earned mention becomes training and retrieval data the next model reads back to a buyer.

At the same time, go deep on your core topic rather than wide on keywords, so the model sees consistent coverage across a whole cluster of related questions instead of a single thin page. Pair that with a clean, consistent entity, the same name, category and facts everywhere, so every mention you earn actually attaches to you. This combination sits at the heart of a serious generative engine optimisation programme, and it is what a proper answer engine optimisation effort is built to deliver.

The order matters. Fix your entity clarity first so nothing you earn leaks away, then build mentions and co-mentions, then deepen your topical coverage to hold the position once you have it. If the picture feels overwhelming, a structured AI visibility audit maps exactly where your competitor is winning and where your fastest gains are, and our case study on a business that closed this gap shows the arc from invisible to cited.

How to know if it is working

Measure the gap directly, not vanity signals. Track how often each assistant names you versus your chosen competitor across a fixed set of buyer questions, and watch that share move over time rather than reacting to any single answer. As your mentions and co-mentions grow, your share of the citations should climb, and it will usually move before your enquiry numbers do, which makes it an early read on whether the work is landing. Run the check across Gemini as well as the others, since each assistant draws on different sources and a gain in one often leads the rest.

Watch your branded search demand too, since it is one of the cleaner signals that recognition is building in the market rather than just on your own pages. When people start searching for you by name more often, the models tend to follow. If you are rebuilding from a standing start, our guide on winning back visibility lost to AI search is a useful companion. The point to hold onto is that your competitor's lead is earned, not permanent, and everything that earned it is something you can earn too.

‍