Answer engine

In full

Answer engines combine a retrieval layer over web or licensed content with a language model that composes the response. From a publisher's perspective they change the unit of visibility from position to citation, compress many results into a few sources, and remove most of the click-through that funded the underlying content. They also differ from each other in index, crawler and citation behaviour, so visibility in one says little about the others.

Example

The same question yields three sources in one assistant and a largely non-overlapping set of five in another.

Related terms

Answer Engine Optimization

Optimising content to be the source of a direct answer in features and assistants that answer questions rather than list links.

Generative Engine Optimization

The practice of improving how often and how accurately a brand's content is retrieved, cited and represented in AI-generated answers.

Citation share

The proportion of AI answer citations across a tracked prompt set that point to a given domain, used as the AI-era analogue of ranking…

Zero-click search

A search where the user's need is satisfied on the results page itself — by a snippet, knowledge panel, calculator or AI Overview — and no…

PerplexityBot

Perplexity's crawler for building its search index, alongside a separate user-triggered agent that fetches pages in response to specific…

Where to read more

63% of Websites Receive AI Traffic (Study of 3,000 Sites) · Study · ahrefs.com
Conductor Academy · Course · conductor.com
Dense Passage Retrieval for Open-Domain Question Answering · Landmark Post · arxiv.org
HubSpot Academy SEO Courses · Certification · academy.hubspot.com
Perplexity Crawlers Documentation · Documentation · docs.perplexity.ai
WebGPT: Browser-assisted Question-answering with Human Feedback · Landmark Post · arxiv.org