Why an assistant names your competitor and not you

What do you and your competitor already share?

Eligibility, usually. The conditions the vendors document are the same for both firms. Google says a page must be indexed and eligible to be shown in Search with a snippet to appear as a supporting link in AI Overviews or AI Mode, that there are no additional technical requirements, and that meeting them does not oblige Google to crawl, index or serve the page. OpenAI says sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though they can still appear as navigational links. Perplexity says PerplexityBot is designed to surface and link websites in its results. Anthropic says disabling Claude-SearchBot may reduce a site's visibility and accuracy in search results. Google adds that AI Overviews are shown only where its systems judge them additive, and so often do not appear.

If both firms clear those conditions, the documents have nothing further to say. None of the four explains how one eligible page is preferred to another, and this piece does not pretend to fill that gap. It asks a smaller question: what could an assistant check, by reading, about each firm, and where would the two differ?

What can an assistant check about a firm by reading?

Three things, as we read the documents. The first is the passage. Google's guide to snippets says they are created mainly from page content, chosen to preview the part that best relates to a user's specific search, so different searches produce different snippets. OpenAI's search tool lists opening a page and finding text inside it as separate actions. Both describe a pipeline that works at the level of passages. A page helps if it holds a stretch of text that answers the question as asked, not merely if the page is about the right subject.

The second is identity. Google's Organization structured data page deals with the problem of telling organizations apart: it says some properties are used behind the scenes to distinguish one organization from another, and a sameAs property can point to the firm's profile pages elsewhere. Google's AI features page says no special markup is needed, so read this as the vendor describing a problem, not prescribing a fix. A system that reads pages from across the web must decide that the name on one is the firm on another. A firm with one name and one sentence of description everywhere makes that decision easy. A firm with variants makes it uncertain.

The third is corroboration. Google's page on helpful, reliable, people-first content asks whether a page shows clear sourcing and background about its author or site, and whether someone researching the site would find it well trusted or widely recognized as an authority on its topic. It concerns Google's ranking systems, not the choice of names in an answer. It does show that agreement from outside the firm is the kind of evidence the vendor itself treats as relevant to a source.

These three are our reading, not a documented formula. No vendor says they decide the outcome or says how they are weighed.

Where would two eligible firms differ?

Exhibit 1 sets the checks side by side for two firms. The firms are constructed for illustration and are not records of any real business. Both pass the documented conditions. Every difference between them is something a system could establish by reading. Read down the rows and note where the firms part: that is where a check stops being a formality and starts deciding which supplier is named. Compare each row with what your own pages publish today.

How can one equally able firm win a question?

Take a retail business asking which firms can connect its store stock system to its online shop. The first firm has a long record of this work. Its page is titled with a service category, opens on a view of where commerce is heading and describes its method in the vocabulary of its brand. It trades under a short name on its site and its full legal name elsewhere. The only outside descriptions of it are directory entries, each written once and never revisited.

The second firm is no better at the work. Its page is titled with the problem, opens with a paragraph that says what it connects to what and for whom, and describes the order in which the work is done. A trade body's member page and a published talk describe it in the same terms. An assistant composing an answer has a sentence to lift from the second firm and checks that other sources agree. For the first it has nothing it can lift and nothing it can confirm.

That is how we would expect it to go, and it is an inference from how firms describe themselves, not an observation of any assistant. It suggests that the firm named is often the one that was easier to describe, which is not always the one that is better at the work.

Which of the three can you change this quarter?

They move at different speeds. The passage is the quickest, because it is your own text: a page can be rewritten in a day and tested the next. Identity takes longer, because it is every place the firm describes itself, which includes the site, social profiles, speaker biographies, directory entries, press boilerplate and the legal name on documents that end up online. One person should hold the approved name and a single approved description, as a style guide holds a brand, and every outside party that asks for a description should be given that text.

Corroboration is the slowest, and the only one that cannot be done alone. A trade body, a publication or a client describes a firm when it has a reason to, and the reason is usually work it can point to. The straightforward route is to do work that is worth describing and to make it easy to describe correctly. A firm that tries to manufacture the mentions is working on the wrong problem, and the documents we read give no basis for thinking it would help.

In what order do you look for the gap?

Exhibit 2 gives the order we use. It starts with access because that part is documented and fixable with certainty: a blocked crawler or a page not eligible to be shown ends the inquiry. It moves to the opening of the page, which is cheap to change and easy to test. Only then does it turn to the harder work of making the firm describable in the same way across its own pages and in other people's.

Run each test on more than one day and on more than one assistant before drawing a conclusion. Google's page says its two AI features may differ in the models and techniques behind them, so the links each shows can differ, and the same should be assumed between vendors.

What if the answer did not come from your pages at all?

Then none of the above applies, and nothing the vendors document tells us how often it happens. Anthropic's documentation says Claude answers directly, without searching, when a request draws on stable knowledge, and OpenAI's guide says the model chooses whether to search. An answer given without a search names firms from the model's own recall. OpenAI states that each crawler setting is independent, so a site may allow its search crawler and disallow its training crawler. The pages we read do not say whether training access changes what a model recalls about a firm, and we do not claim it either way.

For the same reason, do not conclude too much from one answer, and do not copy the competitor's page. Two firms with the same description make identity harder for a reader to resolve, not easier.