AI engine
A consumer or enterprise AI system that answers user queries by retrieving and synthesizing information from across the web. The five engines covered in this Index are ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews.
Retrieval
The process by which an AI engine selects sources to inform an answer. Retrieval is not the same as a Google search result — it is selection that shapes the engine's answer text directly.
AI Retrieval Economy
The set of media properties, institutional publishers, vendor research arms, community substrates, and data publishers whose content shapes AI-generated answers. The retrieval economy operates alongside — and increasingly diverges from — the conventional media economy.
Citation Share
The proportion of AI-engine answers in a sector that cite or rely on a given source. Citation Share is the underlying measure the Index estimates.
Retrieval Anchor
A property scored 72 or above on the composite. Retrieval anchors are the primary citation tier for a sector — sources the engines reliably return to.
Cited tier
Properties scored 56–71. Regularly cited in their sector, but not always as the primary source.
Moderate tier
Properties scored 44–55. Surface occasionally; important but not retrieval-defining.
Low-Yield tier
Properties scored below 44. Rarely in regular engine rotation.
GEO (Generative Engine Optimization)
The discipline of building authority within AI retrieval systems. GEO is to AI engines what SEO is to search engines — but the rules are different, and many of the levers operate on entity authority rather than on keywords.
Entity layer
The internal representation an AI engine maintains of brands, publications, products, people, and concepts as connected entities. The entity layer is what the engine actually retrieves from when answering.
Co-citation
When two sources are cited together repeatedly, the engines associate them with each other in the entity layer, raising the authority of both.
Authority compounding
The phenomenon by which sources cited reliably over time gain additional authority — a feedback loop that explains why long-tenured publications and durable URLs outperform newer or refreshed-and-replaced sources.
Structural finding
A recurring retrieval pattern observed across multiple sectors during the research. Each chapter names one structural finding for its sector; the synthesis chapter at the end of the volume gathers them into a framework.
Extractability
How well a source's content can be parsed, attributed, and summarized by an AI engine. Sources with clean HTML, structured schema, named entities, and stable URLs extract more reliably.