Skip to main content

Glossary description audit

Scope: ai-search-glossary/. Analysis only — no descriptions were rewritten. The regeneration rule is not set yet; this file is the raw material for setting it. A mechanical re-truncation with better arithmetic would reproduce the defect at a new length, so the rule gets decided from the worked examples (Section C) first.

Counts

Extraction method (Sections B–D): the “opening definition sentence” is the first sentence of the first paragraph under each page’s ## Definition heading, verbatim. Proposed-description method (Sections C–D, illustration only): if the opening sentence is ≤155 chars, use it verbatim (it already ends at a sentence boundary); otherwise trim to ≤155 ending at the last clause boundary (, ; : — .) at or before char 155, stripping trailing punctuation, no ....

C. Twenty worked examples

Spanning the range — from pages whose opening sentence is already ≤155 (proposed = verbatim) to the worst cases past 300 chars.

1. author-authority.mdx — Author Authority

  • Current (truncated, 160): Author authority is the credibility and expertise attributed to a content creator — used by search engines and AI systems as a signal of content trustworthin…
  • Opening definition sentence (161): Author authority is the credibility and expertise attributed to a content creator — used by search engines and AI systems as a signal of content trustworthiness.
  • Proposed (81): Author authority is the credibility and expertise attributed to a content creator

2. data-sanitation.mdx — Data Sanitation

  • Current (truncated, 160): Data sanitation is the process of auditing and correcting inconsistent, conflicting, or outdated brand information across digital sources before AI systems i…
  • Opening definition sentence (166): Data sanitation is the process of auditing and correcting inconsistent, conflicting, or outdated brand information across digital sources before AI systems ingest it.
  • Proposed (83): Data sanitation is the process of auditing and correcting inconsistent, conflicting

3. structured-snippet.mdx — Structured Snippet

  • Current (truncated, 160): A structured snippet is a type of rich result that displays a table or list of specific attributes about a product, service, or entity — enabled by structure…
  • Opening definition sentence (171): A structured snippet is a type of rich result that displays a table or list of specific attributes about a product, service, or entity — enabled by structured data markup.
  • Proposed (134): A structured snippet is a type of rich result that displays a table or list of specific attributes about a product, service, or entity

4. comment-signal.mdx — Comment Signal

  • Current (truncated, 160): A comment signal is the engagement and content generated in the comments section of a social media post — including questions, answers, additional informatio…
  • Opening definition sentence (179): A comment signal is the engagement and content generated in the comments section of a social media post — including questions, answers, additional information, and user reactions.
  • Proposed (134): A comment signal is the engagement and content generated in the comments section of a social media post — including questions, answers

5. annual-marketing-plan.mdx — Annual Marketing Plan

  • Current (truncated, 160): An annual marketing plan is a documented strategy outlining a company’s marketing objectives, budget allocation, channel mix, campaign calendar, and performa…
  • Opening definition sentence (194): An annual marketing plan is a documented strategy outlining a company’s marketing objectives, budget allocation, channel mix, campaign calendar, and performance benchmarks for a 12-month period.
  • Proposed (143): An annual marketing plan is a documented strategy outlining a company’s marketing objectives, budget allocation, channel mix, campaign calendar

6. document-embedding.mdx — Document Embedding

  • Current (truncated, 160): Document embedding is the process of converting an entire document — as opposed to individual words or sentences — into a single numerical vector that repres…
  • Opening definition sentence (205): Document embedding is the process of converting an entire document — as opposed to individual words or sentences — into a single numerical vector that represents the document’s overall meaning and content.
  • Proposed (112): Document embedding is the process of converting an entire document — as opposed to individual words or sentences

7. social-content-infrastructure.mdx — Social Content Infrastructure

  • Current (truncated, 160): Social content infrastructure is the systematic architecture of a brand’s social media presence — designed to function as a durable retrieval surface rather …
  • Opening definition sentence (216): Social content infrastructure is the systematic architecture of a brand’s social media presence — designed to function as a durable retrieval surface rather than a series of individual posts optimized for engagement.
  • Proposed (95): Social content infrastructure is the systematic architecture of a brand’s social media presence

8. icp.mdx — Ideal Customer Profile (ICP)

  • Current (truncated, 160): An ideal customer profile (ICP) is a detailed description of the type of company or individual most likely to derive maximum value from a product or service …
  • Opening definition sentence (228): An ideal customer profile (ICP) is a detailed description of the type of company or individual most likely to derive maximum value from a product or service — and therefore most likely to become a long-term, high-value customer.
  • Proposed (148): An ideal customer profile (ICP) is a detailed description of the type of company or individual most likely to derive maximum value from a product or ⚠️ no clause boundary ≤155 — ends at a word boundary; needs a human eye

9. expertise-signal.mdx — Expertise Signal

  • Current (truncated, 160): An expertise signal is any indicator — such as author credentials, publication history, structured data, or domain-specific vocabulary — that communicates a …
  • Opening definition sentence (236): An expertise signal is any indicator — such as author credentials, publication history, structured data, or domain-specific vocabulary — that communicates a content creator’s or brand’s domain expertise to search engines and AI systems.
  • Proposed (134): An expertise signal is any indicator — such as author credentials, publication history, structured data, or domain-specific vocabulary

10. entity-salience.mdx — Entity Salience

  • Current (truncated, 160): Entity salience refers to how central or prominent an entity is within a specific document — how much the document is ‘about’ that entity, as determined by h…
  • Opening definition sentence (248): Entity salience refers to how central or prominent an entity is within a specific document — how much the document is “about” that entity, as determined by how frequently, specifically, and contextually the entity is referenced throughout the text.
  • Proposed (137): Entity salience refers to how central or prominent an entity is within a specific document — how much the document is “about” that entity

11. topical-gap.mdx — Topical Gap

  • Current (truncated, 160): A topical gap is a question, subtopic, or related concept within a brand’s claimed domain of expertise that is not addressed by any existing piece of the bra…
  • Opening definition sentence (257): A topical gap is a question, subtopic, or related concept within a brand’s claimed domain of expertise that is not addressed by any existing piece of the brand’s content — creating a gap in topical coverage that competitors or other sources fill by default.
  • Proposed (153): A topical gap is a question, subtopic, or related concept within a brand’s claimed domain of expertise that is not addressed by any existing piece of the ⚠️ no clause boundary ≤155 — ends at a word boundary; needs a human eye

12. performance-baseline.mdx — Performance Baseline

  • Current (truncated, 160): A performance baseline is the documented measurement of a brand’s current marketing performance across key metrics — before any new strategy, campaign, or op…
  • Opening definition sentence (266): A performance baseline is the documented measurement of a brand’s current marketing performance across key metrics — before any new strategy, campaign, or optimization effort begins — establishing the starting point against which future performance will be measured.
  • Proposed (150): A performance baseline is the documented measurement of a brand’s current marketing performance across key metrics — before any new strategy, campaign

13. knowledge-graph-poisoning.mdx — Knowledge Graph Poisoning

  • Current (truncated, 160): Knowledge graph poisoning is the introduction of inaccurate or misleading information into a knowledge graph — through false Wikipedia edits, incorrect Wikid…
  • Opening definition sentence (276): Knowledge graph poisoning is the introduction of inaccurate or misleading information into a knowledge graph — through false Wikipedia edits, incorrect Wikidata entries, or manipulated structured data — with the effect of corrupting an AI system’s representation of an entity.
  • Proposed (140): Knowledge graph poisoning is the introduction of inaccurate or misleading information into a knowledge graph — through false Wikipedia edits

14. ai-brand-ambassador.mdx — AI Brand Ambassador

  • Current (truncated, 160): An AI brand ambassador is a brand’s deliberate strategy of ensuring that AI systems consistently represent, recommend, and characterize the brand positively …
  • Opening definition sentence (287): An AI brand ambassador is a brand’s deliberate strategy of ensuring that AI systems consistently represent, recommend, and characterize the brand positively across relevant queries — treating AI systems as a form of ambient brand advocacy that operates without direct human intervention.
  • Proposed (117): An AI brand ambassador is a brand’s deliberate strategy of ensuring that AI systems consistently represent, recommend

15. video-description-seo.mdx — Video Description SEO

  • Current (truncated, 160): Video description SEO is the practice of writing YouTube, TikTok, and other platform video descriptions to include target keywords, named entities, related t…
  • Opening definition sentence (295): Video description SEO is the practice of writing YouTube, TikTok, and other platform video descriptions to include target keywords, named entities, related topics, and explicit content summaries — optimizing the text field that AI systems use as the primary parseable document for video content.
  • Proposed (146): Video description SEO is the practice of writing YouTube, TikTok, and other platform video descriptions to include target keywords, named entities

16. ai-citation-strategy.mdx — AI Citation Strategy

  • Current (truncated, 160): An AI citation strategy is a deliberate approach to earning references within AI-generated responses — combining content structure, authority signal building…
  • Opening definition sentence (306): An AI citation strategy is a deliberate approach to earning references within AI-generated responses — combining content structure, authority signal building, entity optimization, and multi-platform presence to systematically improve how often and how accurately a brand is cited across AI search surfaces.
  • Proposed (130): An AI citation strategy is a deliberate approach to earning references within AI-generated responses — combining content structure

17. semantic-triple.mdx — Semantic Triple

  • Current (truncated, 160): A semantic triple is a fundamental unit of knowledge representation in the form of subject–predicate–object — for example, ‘Plate Lunch Collective – is locat…
  • Opening definition sentence (315): A semantic triple is a fundamental unit of knowledge representation in the form of subject–predicate–object — for example, “Plate Lunch Collective – is located in – Hawaii.” Semantic triples are the building blocks of knowledge graphs and linked data systems, enabling machines to reason about entity relationships.
  • Proposed (146): A semantic triple is a fundamental unit of knowledge representation in the form of subject–predicate–object — for example, “Plate Lunch Collective

18. short-form-video-seo.mdx — Short-Form Video SEO

  • Current (truncated, 160): Short-form video SEO is the practice of optimizing videos under 60–90 seconds on platforms like TikTok, Instagram Reels, and YouTube Shorts for discovery thr…
  • Opening definition sentence (325): Short-form video SEO is the practice of optimizing videos under 60–90 seconds on platforms like TikTok, Instagram Reels, and YouTube Shorts for discovery through platform search and AI retrieval — using keyword-rich titles, captions, spoken keywords, on-screen text, and hashtags to improve topical clarity and searchability.
  • Proposed (119): Short-form video SEO is the practice of optimizing videos under 60–90 seconds on platforms like TikTok, Instagram Reels

19. ai-agent-discoverability.mdx — AI Agent Discoverability

  • Current (truncated, 160): AI agent discoverability is the degree to which a brand’s content, entity signals, and digital infrastructure are accessible and legible to AI agents — auton…
  • Opening definition sentence (341): AI agent discoverability is the degree to which a brand’s content, entity signals, and digital infrastructure are accessible and legible to AI agents — autonomous systems that conduct research, make recommendations, and take actions on behalf of users — as distinct from human-facing discoverability or single-turn AI search discoverability.
  • Proposed (149): AI agent discoverability is the degree to which a brand’s content, entity signals, and digital infrastructure are accessible and legible to AI agents

20. brand-grounding.mdx — Brand Grounding

  • Current (truncated, 160): Brand grounding is the practice of providing AI systems with accurate, structured, verified information about a brand — through schema markup, Wikidata entri…
  • Opening definition sentence (371): Brand grounding is the practice of providing AI systems with accurate, structured, verified information about a brand — through schema markup, Wikidata entries, authoritative third-party citations, and entity infrastructure — so that AI-generated responses about the brand are anchored in factual data rather than generated from incomplete or inaccurate training signals.
  • Proposed (141): Brand grounding is the practice of providing AI systems with accurate, structured, verified information about a brand — through schema markup

D. Rule-break cases (25)

Pages where “take the opening ## Definition sentence and trim to ≤155 at a clause boundary” does not cleanly apply. These need a human decision, not the mechanical rule.

E. Untruncated outliers (18) — 161–185 chars, no ...

Separate question from the 340. Not mechanically truncated; several read as hand-rephrased (e.g. parametric-belief). May or may not need trimming for snippet length — flagged for a separate decision.

A + B. Full inventory with source sentences (340)

Every affected file: path, title, current truncated description + char count (A), and the verbatim opening ## Definition sentence + char count (B). Combined into one table for scannability.