How often AI cites Netflix

Netflix's domain holds three kinds of pages that AI models can use:

1. Title pages, one for each film or series, with a synopsis, cast, genres and age rating. They confirm that a title exists on the service and describe it in a few lines.
2. Help Center articles, which explain subscription plans, payment, profiles, downloads, account sharing rules and supported devices.
3. Corporate and editorial pages, such as the company newsroom, its careers and culture pages, and Tudum, Netflix's official site about its shows.

None of these pages offers independent opinions. For "is this series worth watching", assistants turn to critics, review aggregators and discussion forums; Netflix itself serves to confirm availability and basic facts.

Citations follow consumer questions about the service itself or its catalog:

• "How much does Netflix cost per month" or "which Netflix plan has no ads"
• "How to cancel Netflix" or "how to add an extra member"
• "Is [series] on Netflix" or "what is [film] about"
• "Best documentaries about entrepreneurship on Netflix"

Business prompts can also mention Netflix without linking to it: the company is a classic example in answers about subscription models, recommendation algorithms or company culture. In those cases, AI models usually cite case studies and business media, or Netflix's own corporate texts, rather than the consumer site.

Netflix is a localized service: the catalog, prices and plans differ from one country to another. Assistants that search live, like Perplexity, ChatGPT with search or Gemini (grounded in Google Search), can retrieve the version that matches the user's country and quote the help page that applies.

Models answering without search may give outdated prices or availability. A model that does not search, as Claude often does by default, tends to recommend checking the official site rather than link to a page.

Grok has a different angle: release announcements and reactions to new seasons circulate quickly on X, so it can answer "when does season 3 come out" from posts rather than from netflix.com. These different routes explain why one model may link to Netflix while another answers the same question without it.

The questions behind this dataset concern companies, their products and their markets, not evening entertainment. Netflix figures are therefore small and can swing from one month to the next on a handful of answers. A spike usually reflects a few questions about streaming, media or subscription pricing, not a shift in how AI models see the service.

The figure also counts links, not mentions. An answer that names Netflix as an example of a subscription business without linking to netflix.com does not add to it.