AI competitor analysis: who AI names next to your brand

Every answer that names two brands together is a small vote that they belong in the same set. Aggregate a month of those votes and you get a competitive map drawn by the model, not by your marketing team.

In a single 30-day window that map held 153,572 distinct brand pairs and 1.3 million co-mentions across 29,797 brands. It is detailed enough to read your own category from.

The single brand AI names alongside yours most often is a brand that names you back as its own number one in fewer than one in four cases. Most of the time, the brand you are most tied to barely registers you.

Even when your top partner does acknowledge you, it is rarely near the top. It places you in its own top three only 40.8% of the time, and in its top five only 51.6% of the time.

Among brands with enough pairings to measure, the typical brand ranks fifth on the list of the very brand it is most associated with. You point at them first; they point back, if at all, from the middle of the pack.

Group the active brands by what their single strongest pairing actually is and three outcomes appear. Three in four ride the coattails of a bigger brand that does not name them back. Only a quarter are genuine mutual rivalries, and just 1.7% are the reverse.

The reason coattails dominate is size. In 85.2% of strongest pairings the partner is the larger brand by co-mention volume. Attention concentrates upward, so the smaller brand is tied tightly to a giant for whom it is a minor association.

It is not a near-peer relationship. The median strongest partner carries roughly three times the co-mention volume of the brand tied to it. A buyer asking about them hears you in passing; a buyer asking about you hears them as a main comparison.

When the pairing is not mutual, the gap is stark: 44.2% of the time the bigger partner does not place you anywhere in its top ten, and the median non-mutual rank is ninth. They are firmly on your radar; you are barely on theirs.

The one-sided shape is clearest in crowded software categories. AI names the bigger brand beside the smaller one constantly, while the bigger brand ranks the smaller one far down its own list.

The roughly one in four pairings that go both ways are where AI and the market agree on a head-to-head. These read like rivalries a category analyst would draw by hand.

Separate rivalries from coattails. Beating a coattails brand on a feature grid will not change how AI pairs you; that pairing is context, not a contest.

Look past the top partner. Ranks two through ten are where competitors from adjacent categories surface, along with the ones that never make your own slide.

Watch the shape over time. A strengthening coattails link, or a newly forming mutual one, is visible weeks before win rates move.

We counted how often each pair of brands is named together inside the same AI answer, across ChatGPT Search and Google AI Mode, over the 30 days ending June 26, 2026. For each brand we took its most co-mentioned partner, then looked up where the brand ranks on that partner's own list. A pairing is mutual when each brand is the other's number one.

The active set is limited to brands with enough distinct pairings to be stable, about 7,250 brands. Co-mention measures association, not a recommendation verdict or sentiment: two brands named together may be framed as alternatives, complements, or one as a foil for the other. Figures describe an observed sample of AI answers from the two engines in scope, not Perplexity, Gemini, or Copilot.