How AI engines choose which brands to recommend

Ask ChatGPT to recommend a vendor in your category and it will name three or four companies. The list is rarely the four biggest. It is often not the four that rank highest in Google either. Understanding why is the difference between guessing at AI visibility and working on it.

Five signals do most of the deciding.

1. Can it tell what you are?

Before an engine can recommend you for something, it needs a confident answer to what you are. This sounds trivial and it is where most brands lose. Positioning written to sound impressive reads as noise: "we unlock growth through intelligent experiences" gives a model nothing to match a query against.

Clear category language on your own site, consistent Organization and Service schema, and the same name, address and description everywhere you appear. Boring, and it is the foundation everything else sits on.

2. Do other people say the same thing about you?

This is the signal that most reliably explains the surprising winners. Engines corroborate across sources. A claim that appears only on your own domain is weak; the same claim appearing on review platforms, comparison articles and industry publications becomes fact as far as the model is concerned.

It is also why a smaller, well-covered competitor beats a larger, quieter one. Size does not propagate. Coverage does.

3. Can it lift a clean answer off your page?

Generative engines compose answers from fragments they can extract. That rewards a specific shape of writing: a question-shaped heading, a direct and self-contained answer in the first two sentences, then the depth below.

The same information written as flowing prose, with the answer arriving in paragraph six after the context, is functionally invisible. Not because it is worse for humans, but because there is no clean fragment to quote. This is the single most common fixable problem we find.

4. Can it reach you at all?

Two hard gates, both binary, both silent when they fail:

• Crawler access. If OAI-SearchBot, PerplexityBot, Google-Extended or ClaudeBot are blocked, you are opted out. Check the CDN layer too, where managed rules can block AI crawlers without appearing in the robots.txt you wrote.
• Index coverage. ChatGPT searches the live web through Bing. A site indexed thoroughly in Google and barely in Bing is invisible to ChatGPT's browsing path, which is a common and largely unnoticed failure.

5. Is your information current?

Engines prefer recent sources for anything that could have changed: pricing, features, comparisons, regulation. A comprehensive page last updated three years ago loses to a thinner page updated last month. Dates in schema, dates visible on the page, and genuine updates rather than a changed timestamp.

Why the biggest brand often loses

Put those five together and the pattern makes sense. Large incumbents frequently have vague marketing-led positioning, sprawling sites where answers are buried, aggressive bot management that blocks AI crawlers, and content that has not been touched in years. A focused competitor with clear language, a clean structure, open crawlers and active third-party coverage outperforms them inside AI answers while losing badly on every traditional authority metric.

That gap is temporary. It is also the reason to work on this now rather than when the category has caught up.

How to find out where you stand

There is no analytics dashboard for this. The only real method is to ask the questions your customers ask, across every engine, on a schedule, and record what comes back: who gets named, in what order, with what sentiment, citing which sources. That last part is the most actionable output, because it tells you exactly which third-party pages to work on instead of guessing.

Our free AI visibility checker gives you a baseline in about a minute. For the tactics behind each signal, see how to rank in ChatGPT, and for the strategic frame, what Generative Engine Optimization is.