AI Search Academy: Glossary

Every low visibility score traces to one of three gaps: SEO Gap, PR Gap, Content Gap. Fix in that order.

Minimum viable prompt set: 3 intents × 2 scopes × 2 visibility-hygiene states = 12 tracked prompts per segment.

Answer Engine Optimization. Older term, largely subsumed by GEO.

Your URL is linked as a source inside an AI answer. Signals trust.

Your brand name appears in the AI answer text without a linked citation. Signals awareness.

The shift from keyword + 10 blue links to natural-language question + one synthesized answer, with brands competing to be cited inside the source pool.

The post-click-economy reality: 60% of searches end without a click, so you compete to be in the answer, not for a blue link.

Bottom Line Up Front. A 40-80 word direct answer in the first paragraph. The single highest-leverage content pattern.

A prompt that contains your brand name. Tests visibility hygiene: does AI know you exist and describe you correctly?

A paragraph-or-heading-bounded unit that RAG systems extract, embed, and quote. The unit AI engines actually cite - not pages.

How easily an LLM can lift a self-contained answer block from your page. Driven by 50-120 word self-contained paragraphs, clear headers, BLUF.

The pre-AI-search world of keyword → 10 blue links → click-through. Collapsing now.

Third of the 3 Gaps. Either content on the topic does not exist, or it exists but AI engines cannot extract a clean answer. Four patterns: BLUF, Justification, Freshness, Evidence.

% of prompt runs where your brand appears at all (mention OR citation). Branded detection below 90% = entity emergency.

A named thing an AI engine recognises: company, product, person, place, concept. GEO is partly the fight to become a recognised entity.

Generative Engine Optimization. Optimising to be visible, cited, and recommended inside AI-generated answers.

One canonical URL AI engines reach for ground-truth facts about your brand. The anchor for branded visibility.

Prompts carry one of three intents: Information (learning), Consideration (comparing), Transaction (buying). One of the three required tag dimensions.

The schema format AI engines prefer. A <script type="application/ld+json"> block declaring entities, products, FAQs in machine-readable form.

Pros/cons, comparison tables, "best for X" statements. What engines extract when deciding who to recommend.

Google's structured entity card. Strong signal you are an established entity; wrong panel data spreads faster than right data.

Bot that fetches pages for LLM training or live retrieval. Must be allowed: GPTBot, CCBot, PerplexityBot, ClaudeBot.

A prompt without your brand name. Tests whether AI associates you with the category. Where the real competitive battle is.

Rankscale feature. Audits a single URL against AI-engine readiness criteria (crawlability, schema, hierarchy, chunkability).

Where you show up in the answer: #1-3 strong, #4-6 mid, #7+ weak, Not Found.

Second of the 3 Gaps. Third-party sources do not know, cite, or describe your brand the way they cite competitors. Three sub-types.

An AI engine that generates answers by predicting likely tokens from training data + retrieval, not by matching keywords. Why each extra word shifts the output.

A thematic bundle of prompts around one topic, job-to-be-done, or funnel stage. The "what we're tracking."

A structured set of prompts varying across intent / scope / visibility-hygiene. Minimum viable: 3x2x2.

Different formulations of the same prompt group to sample performance across buying journey phases, branded vs non-branded, and segment vs niche.

Rankscale feature. Auto-generates per-prompt, per-engine recommendations and content ideas from the Sentiment Radar.

Retrieval-Augmented Generation. The mechanism by which AI engines retrieve relevant chunks and synthesise answers. Explains why chunkability matters.

The vocabulary JSON-LD uses. Defines Organization, Article, Product, FAQPage, HowTo, and more.

Second required tag dimension: Segment vs Niche in segment vs Niche-in-niche vs Geography. Narrower scope = easier to win first.

One H1, 3-8 question-format H2s, scannable H3s, no skipped levels. The structure RAG uses to define chunk boundaries.

How AI describes you: positive, neutral, mixed, negative. Below 55% positive = reputation risk.

First of the 3 Gaps. Technical or structural issues blocking AI engines from reaching, parsing, or chunking your pages. Fix first; nothing else compounds without it.

Your % of citations in a prompt group vs all competitors. Headline metric for the "We're Defending" narrative.

Your % of mentions + citations vs competitors. Broader than Share of Citations. Use this as a market benchmark.

Labels on every prompt variation. Three required (visibility hygiene, intent, scope); optional extensions. No tags = no action possible later.

The gap between being mentioned and being cited. AI knows your name but will not source your pages.

First required tag dimension: Branded vs Non-Branded. Branded is the floor; non-branded is the ceiling.

Your aggregate 0-100 score on a prompt variation or prompt group. Headline metric. Distinguish between branded & non-branded prompts. Hides the diagnosis unless paired with the other 6.