Getting named by AI: how generative search chooses

Search is moving from a list of links to a single answer. Assistants such as ChatGPT, Perplexity and Gemini, and the AI overviews in Google itself, summarise and name a few businesses. The customer clicks less and trusts the answer more.

That changes the question for companies. Not only: do I rank high? But: does the answer name me, and why or why not?

How an answer comes about

An AI answer is not chance. The model retrieves sources, weighs them and composes an answer. Which companies get named depends on signals you largely influence yourself.

1. 01ClarityThe model must understand who you are, what you do and where. The same facts on your site, your business profile and other sources make you recognisable.
2. 02Outside proofWhat others say about you weighs more than what you say about yourself: reviews, mentions, trade publications.
3. 03Answers to real questionsPages that answer a concrete question directly are used as a source more often than general copy.
4. 04FreshnessRecent, maintained information beats pages nobody has touched in years.

Measuring what you cannot see

The hard part: you do not notice when an assistant does not name you. No position drops, no click goes missing in your report. So you have to measure it actively. Regularly ask the questions your customers ask, across several assistants, and track who gets named and which sources that rests on.

Visibility in AI is not a position but a mention. And a mention has to be earned with proof.

What this means

1. 1Map the twenty questions your customers ask before they choose. Use their words, not yours.
2. 2Measure per question whether you are named, in which assistant and next to whom.
3. 3For every question where you are missing, create a page that answers it directly, with the facts a model needs.
4. 4Treat mentions by others as part of your marketing, not as an afterthought.