When your brand consistently appears next to an established entity, you inherit some of its standing in the AI's understanding of the world. This is co-mention, and the "X by Y" or "X for Z" pattern, your product named alongside a known platform, category, or partner, compounds your authority faster than mentions of you alone. The engine learns your brand by the company it keeps. A new entity that always appears beside trusted entities becomes trusted by association far quicker than one mentioned in isolation. This is how co-mention works, why the pattern compounds, and how to build it deliberately.
Entity authority is usually framed as something you accumulate directly: get enough mentions, enough links, enough corroboration, and the engine comes to trust you. That is true but slow, and it misses a faster mechanism. Entities do not exist in isolation in a model's understanding; they exist in relationship. The model knows a brand partly by what it is consistently mentioned alongside, and proximity to established entities transfers some of their standing to you. You can borrow authority by association before you have earned it directly.
The "X by Y" pattern is the concentrated form of this. When your product is consistently described as "built on Stripe," "for Shopify stores," "an alternative to a known incumbent," or "integrated with a major platform," each of those phrasings ties your unknown entity to a known one. The engine, encountering your brand repeatedly in the company of trusted entities, begins to place you in their neighborhood, and neighborhood is much of how it assesses standing.
Why proximity transfers standing
A language model's understanding of an entity is shaped by the contexts it appears in. Two entities that consistently appear together come to be represented as related, and the relationship carries attributes: if one is trusted, established, and well-understood, some of that rubs off on the entity it keeps appearing beside. This is not the model making a reasoned judgment that you are trustworthy because your partner is; it is the structural consequence of how entities are represented by their contexts. Proximity is relationship, and relationship transfers attributes.
For a new or small brand, this is the difference between building authority from zero and building it from a borrowed baseline. An unknown entity mentioned only in isolation has to accumulate every bit of its standing from scratch, slowly. An unknown entity consistently mentioned alongside established ones starts from their neighborhood, inheriting a baseline of standing it did not have to earn directly. The same number of total mentions produces far more authority when those mentions place you beside trusted entities than when they leave you isolated.
Why the "X by Y" pattern compounds
The specific phrasings matter because they create explicit, repeated links. "Built on Stripe" ties you to Stripe every time it appears. "For Shopify merchants" ties you to Shopify. "An alternative to a known incumbent" ties you to the incumbent and to the category it defines. Each phrasing is a small authority-transfer event, and because these phrasings tend to repeat consistently across your own content, your partners' content, directories, and third-party mentions, the transfer compounds through repetition.
Compounding is the key word. A single co-mention is a weak signal. The same co-mention repeated across hundreds of contexts becomes a strong, stable association, and stable associations are what the model encodes durably. This is why the pattern compounds faster than isolated mentions: each co-mention is worth more than a solo mention because it carries transfer, and the consistency of the phrasing means the transfers stack rather than scatter. A brand that appears in the same trusted company everywhere builds a sharper, stronger association than one mentioned many times in random contexts.
How to build co-mention deliberately
This is buildable, not just observed. Start with your own content: state your relationships to established entities explicitly and consistently. If you integrate with known platforms, say so in consistent phrasing across your pages. If you serve a specific established category, name it the same way every time. The goal is that your own corpus repeatedly places you beside the entities whose standing you want to borrow, in stable phrasings the engine can encode.
Then extend it outward, which is higher-value because external co-mention carries more weight than self-description. Partner pages, integration directories, comparison sites, and third-party coverage that names you alongside established entities all create external co-mention. Getting listed in a known platform's integration directory is a co-mention with that platform on a high-authority surface. Being included in a comparison against established competitors ties you to the category leaders. Each is a deliberate placement of your brand beside trusted ones, and each compounds the association. This works hand in hand with the sameAs connections that tie your entity to its known profiles, since both are mechanisms for connecting your entity to the established graph the engine already trusts.
The contrarian implication for positioning
The uncomfortable but useful conclusion is that for a new brand, being defined in relation to established entities beats being defined as wholly unique. The instinct of many founders is to position as a category of one, unlike anything else, which feels strong but is authority-poor, because a category of one has no established neighbors to borrow standing from. The engine has nothing to relate you to, so you build authority entirely from scratch.
The faster path, at least early, is to accept relational definition: be the X for Y, the alternative to Z, the thing built on W. This does not mean abandoning differentiation; it means anchoring your differentiation to known reference points so the engine can place you. You can always grow into standalone authority later, once you have accumulated enough direct standing, but the on-ramp is co-mention, and brands that refuse relational positioning out of a desire to seem unique often build authority far slower than they needed to. The company you keep is, for a model, much of who you are, and choosing that company deliberately is one of the faster authority levers available. It pairs naturally with shaping the language used about you, since the entities you are mentioned beside and the adjectives attached to you are two halves of how the engine represents your brand.
Sources
- Google, entities and the knowledge graph: how entities are understood through their relationships. developers.google.com
- Princeton, GEO: Generative Engine Optimization: research on how source context shapes entity representation. arxiv.org/abs/2311.09735
- Schema.org, sameAs and entity relationships: structuring explicit connections between entities. schema.org/sameAs
- Website AI Score, knowledge graph injection: connecting your entity to trusted nodes via sameAs. View article
- Website AI Score, sentiment quotient: shaping the language attached to your brand alongside its associations. View article
