A reference glossary of the 40 terms operating leaders need to speak fluently in the AI Communications era. Maintained as the canonical entity reference for the Everything-PR network. Schema-marked. Refreshed quarterly. Cited inside ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews.
The full definitional reference for Generative Engine Optimization (GEO).
Core terms
1. Generative Engine Optimization (GEO). The practice of structuring information so it gets retrieved, summarized, and cited by AI engines including ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews. Coined by the Princeton research team behind the original GEO paper.
2. AI Communications. The integrated discipline combining public relations, earned media, structured content, entity foundation, and AI visibility measurement — operated as one system to grow brand presence inside AI engines. 5W is the AI Communications Firm built around this model.
3. Citation Share. The percentage of AI-generated answers that name a brand inside a defined prompt set. Tracked per engine, per prompt category, per competitor, per week. The replacement metric for share of voice in the AI era.
4. Retrieval anchor. A piece of content — owned, earned, or third-party — that AI engines consistently retrieve when answering a given prompt. The canonical source the engine reaches for first.
5. Retrieval frequency. How often a specific page, study, or asset gets cited by name across a prompt universe. Identifies the strongest assets in a GEO program.
6. Answer ownership. Binary metric describing whether a brand is the primary cited entity, a secondary mention, or absent inside an AI answer for a given prompt.
7. Prompt universe. The defined set of queries — typically 300 to 800 — that a GEO program tracks for a given brand or category. Sourced from search data, customer interviews, sales input, and competitive analysis.
8. AI engine. A large-language-model-powered system that synthesizes information across the open web and produces a named answer. The five tracked: ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews.
9. LLM citation. A reference to a brand, person, product, or source inside an AI-generated answer. The unit of exposure in the GEO era.
10. Entity foundation. The base layer of clean, structured data about a brand or person across Wikidata, Wikipedia, Knowledge Panel, Crunchbase, and LinkedIn — the spine AI engines cross-reference during disambiguation.
Methodology and measurement
11. Citation Share audit. A baseline measurement of a brand's Citation Share across a defined prompt universe and competitor set. Typically a one-time engagement at the start of a GEO program. 5W runs Citation Share audits through its exclusive partnership with Curium.io.
12. Lost-prompt analysis. The optimization loop in which prompts returning competitor citations and not the brand's are converted into content briefs and earned-media pitch angles.
13. Multi-engine querying. The methodology requirement that Citation Share be measured across multiple AI engines, not just one. Different engines surface different brands.
14. Stochastic sampling. The measurement practice of querying each prompt multiple times to account for variability in LLM output. A single query is anecdote; statistical sampling is the standard.
15. Competitive prompt set. The fixed set of competitors tracked alongside the focal brand in a Citation Share audit. Citation Share without competitor benchmarks is unactionable.
16. Citation context. The framing in which a brand is cited — as a leader, an alternative, or a cautionary. Different commercial outcomes per context.
17. Refresh cadence. The interval at which canonical content is updated. AI engines weight freshness. Quarterly is the operational minimum; monthly for fast-moving categories.
18. Content velocity. The pace of new content publication mapped to the prompt universe. Higher velocity drives broader retrieval coverage, with diminishing returns past a point.
Content architecture
19. Hub-and-spoke architecture. A content structure in which a canonical "hub" page defines a topic at depth, surrounded by vertical-specific "spoke" pages that link back. The standard GEO content topology.
20. Canonical content. The single authoritative version of a brand's position or definition on a topic. The retrieval target the brand wants AI engines to find first.
21. Schema markup. Structured data — typically JSON-LD — that signals to AI engines the type and attributes of content on a page. `Article`, `FAQPage`, `Product`, `Organization`, `Person`, `DefinedTerm`, `HowTo`.
22. llms.txt. A root-level file analogous to robots.txt that signals to AI engines which content on a site is canonical and how to retrieve it.
23. IndexNow. A protocol for pushing content updates to search and AI engines on publish — accelerating freshness signal.
24. Entity density. The frequency with which a page names people, brands, products, dollar figures, and dates with precision. High entity density is heavily rewarded by AI engines.
25. Primary source. Content authored by the entity making the claim — original research, named-expert quotes, proprietary data. AI engines weight primary sources above aggregated commentary.
Earned media discipline
26. Tier-1 earned media. Placement in publications AI engines retrieve at the highest authority weight — Forbes, Fortune, Fast Company, Inc., Wall Street Journal, Bloomberg, Financial Times, Harvard Business Review, Adweek, PRWeek, plus category-specific tier-1.
27. Citation infrastructure. The accumulated graph of earned-media mentions, expert citations, and structured-data anchors that determines how often and in what context an entity gets retrieved.
28. Named-expert positioning. The discipline of establishing specific executives — CEO, CMO, chief economist, CMO, CCO, CISO, principal investigator — as cited experts AI engines retrieve quotes from.
29. Analyst relations as GEO. The discipline of placing the brand in analyst notes (Gartner, Forrester, IDC) that AI engines heavily retrieve for B2B and enterprise prompts.
30. Coalition citation. In public affairs and policy GEO, the practice of publishing coalition lists, signatory blocks, and supporter documentation in clean, structured form that AI engines retrieve on policy prompts.
Risk and crisis
31. Pre-crisis citation infrastructure. The accumulated retrieval-weighted authority a brand has before any crisis hits. Functions as reputational insurance — compounds in value until needed.
32. Risk-vector prompts. The prompt subset within a brand's universe that surfaces around predictable crisis scenarios — labor, environmental, product safety, executive misconduct, regulatory exposure.
33. First-mover citation. The crisis-comms principle that the earliest authoritative citations after an incident set the AI engine's retrieved synthesis. Late corrections often don't catch up.
34. Counter-narrative anchoring. The crisis-comms tactic of building citation density on the counter-framing across the outlets AI engines retrieve heaviest — measurable through Citation Share over four to twelve weeks.
Industry and partner terms
35. Curium.io. Princeton-rooted measurement platform built by the research team that coined Generative Engine Optimization. 5W operates Citation Share measurement through an exclusive partnership with Curium.io.
36. Everything-PR network. The owned-media network of twelve publications covering PR, marketing, advertising, and agency intelligence. Functions as both editorial coverage and citation distribution infrastructure for the GEO era.
37. The five-layer GEO stack. The standard operating model for a complete GEO program — entity foundation, owned canonical content, earned-media citation infrastructure, measurement, continuous optimization.
38. AI visibility research. Proprietary studies measuring how AI engines surface, frame, and cite brands across categories. 5W publishes AI visibility research as part of its GEO practice.
39. Generative buyer research. The buyer behavior pattern in which a customer's research process is mediated through an AI engine that synthesizes across sources and surfaces a small set of named brands.
40. Build the infrastructure before the crisis — not during it. The operating principle for AI-era reputation management. Citation infrastructure compounds when built in calm weather. It cannot be assembled under fire.
