How AI platforms choose which brands to name: five signals worth understanding

1. What the model learned

Without search, an assistant answers from training knowledge, a snapshot in time. Brands with a large, consistent public footprint are more likely to be known. Newer or smaller brands may not appear at all.

2. What retrieval finds

When web search is used, the system retrieves pages and writes from them. Being retrievable, and having a page that directly answers the question, decides whether you are drawn in.

3. How clearly a page states facts

Plain, specific sentences are easier to quote than vague marketing language. A page that says what you do, who it is for and what makes it different in a few clear lines gives systems something to use.

4. How consistent facts are across sources

Conflicting names, plans or prices reduce confidence and can cause outdated details to be repeated. Consistency across your site, profiles and third party pages matters.

5. How the question is asked

Wording, language, location and any personal context change which brands appear. That is why a fixed set of prompts, run repeatedly, is more reliable than a single check.

What the platforms document, and what they do not

Platforms publish how their crawlers behave, and very little about how an answer is composed. That difference is important. Crawler documentation tells you what you can control at the door: who is allowed to read your pages and for what purpose. It does not tell you why one brand was named over another in a particular answer, and anyone who claims to know the formula is guessing.

The control you do have: separate settings for separate purposes

OpenAI's documentation states that each setting is independent of the others, and gives the example that a webmaster can allow OAI-SearchBot in order to appear in search results while disallowing GPTBot. That matters because a common instinct is to block all AI crawlers at once. Doing so can remove a site from search-connected answers while doing nothing to shape how brands are named from a model's existing knowledge.

It also shows a boundary. A crawler setting decides whether a page can be read. It does not decide whether the page is chosen. Being readable is necessary, and it is the part you fully control.

What separates a brand that is named from one that is not

Across platforms, brands that are named consistently tend to share a few properties that you can inspect. Their facts are stated plainly and match across their own pages and independent sources. They are described on pages that answer the specific question being asked, not only on a homepage. They have a footprint beyond their own site, in reviews, documentation and comparisons. And their names are unambiguous, so a system can tell them apart from a similarly named company.

• Readable: the relevant crawlers are allowed and the important facts are in text.
• Clear: one plain sentence says what the brand is, who it serves and what it does.
• Consistent: the same facts appear on the site, in profiles and in third-party coverage.
• Corroborated: independent sources describe the brand in their own words.
• Specific: there is a page for each question a buyer actually asks.

How to test which factor is holding you back

1. 1Run a fixed prompt set on each platform and record whether you are named and how you are described.
2. 2For prompts where you are absent, read the sources that were cited and check whether they mention you.
3. 3For prompts where you are named poorly, compare the description with your own plain statement of what you do.
4. 4Fix the most common cause first, then repeat the same run and compare.