An evidence-led reference to the concepts, metrics and mechanisms that determine how brands are discovered, cited and recommended by AI systems. Each entry establishes a precise definition, explains how the concept is measured or how it works, and connects it to the methods and research behind it.

Scope

The Index covers AI-driven search and answer surfaces: ChatGPT search, Perplexity, Google AI Overviews and AI Mode, Gemini, and Claude. Entries are written and maintained by Harpal Singh and reviewed on a rolling basis, with a visible last-reviewed date on every page. Metrics defined by blimpp are labelled as such and published with their measurement method; no proprietary term is presented as an established industry standard.

Featured concepts

Browse by category

Measurement. The metrics used to quantify AI visibility: AI visibilitycitation probabilitycitation volatilityrecommendation shareAI share of voicebrand mention probabilityprompt coveragecross-engine citation overlap.

Retrieval. How systems discover and select information: query fan-outgroundingretrieval-augmented generationpassage retrievalsource selectioncitation absorptionextractability.

Entities. Identity and association signals: entity associationentity saliencebrand disambiguationattribute associationentity homeknowledge graph presence.

Authority. External evidence and consensus: source influencethird-party consensusReddit influencepublisher influenceearned authorityreview influenceinformation gain.

Platforms. Stable references for the major AI search surfaces: ChatGPT searchPerplexityGeminiGoogle AI ModeGoogle AI OverviewsClaude, and the AI search crawlers.

Optimisation. Industry practice and intervention types: AI SEOgenerative engine optimisationanswer engine optimisationretrieval optimisationcitation engineeringAI reputation managementllms.txt.

Methodology and research

Every metric in the Index links to a methodology page describing how it is measured, including sampling, repetition and recording conditions. Start with the repeated-run protocol and how citations, mentions and recommendations are measured. The AI Search Research programme sets out the studies planned against these methods and the evidence standards they follow.

About the Index

The Index is published by blimpp and edited by Harpal Singh. Entries are corrected openly: material changes update the visible review date, and corrections are noted on the page they affect. Definitions are written to be self-contained and non-promotional so they can be referenced accurately by other writers and systems. To see how your own brand currently appears across AI engines, you can run a free check at leoseo.com.

Editor: Harpal Singh · Last reviewed: 7 August 2026