AI Visibility & GEO glossary

Plain-language definitions of the terms behind AI visibility and Generative Engine Optimization, the vocabulary AI engines and their optimizers actually use in 2026.

AI Visibility

How present, cited and recommended a brand is inside the answers generated by AI engines (ChatGPT, Perplexity, Gemini, Google AI Overviews, Copilot). The discipline of improving it is called GEO.

GEO (Generative Engine Optimization)

The practice of getting a brand named and cited inside AI-generated answers. GEO targets the citation slot inside the answer, where SEO targets a rank in the list of links.

AEO (Answer Engine Optimization)

Optimizing specifically to win the answer an engine returns. Closely related to GEO and often used interchangeably.

SCR™ Score

A 0–100 measure of how often AI recommends a brand: presence (0.6) + citation (0.4), measured separately across five engines. Maps to three rungs: Surfaced, Cited, Recommended.

Surfaced

The first SCR rung: whether an engine names or mentions the brand at all when its category is asked about. Raw presence.

Cited

The second SCR rung: whether, when surfaced, the brand is quoted as a referenced source, and represented accurately, not misclassified.

Recommended

The third SCR rung: whether the brand is presented as the preferred option versus competitors named in the same answer. Where deals are won.

Presence

The share of relevant AI answers in which a brand is named. One of the two inputs to the SCR™ Score.

Citation

The share of source-exposing AI answers (e.g. Perplexity, Google AI Overviews) that reference a brand's domain as a source.

Answer-first block

A short (40–60 word) self-contained answer placed at the top of a page, so an engine can lift it directly. Around 44% of AI citations come from the first 30% of a document.

Entity

The machine-readable identity of a brand, what it is, who it serves, the problem it solves, that lets engines confidently classify and recommend it. Strengthened with schema, consistent naming and knowledge-graph links.

Entity co-occurrence

How often a brand's name appears alongside category leaders in third-party content. Strong co-occurrence teaches the model which category the brand belongs to.

AI Overviews

Google's AI-generated answer block shown above traditional results. FAQPage schema notably improves the odds of being cited in it.

Narrative Codex

The canonical set of words, claims and terminology a brand wants AI to repeat, the language seeded across its own pages and earned media so the model tells the right story.

Earned media / off-site sources

Third-party places a brand is mentioned, listicles, Reddit, YouTube, LinkedIn, reviews. Roughly 80–90% of brand mentions in AI answers come from off-site sources.

llms.txt

A proposed machine-readable file at a site's root. Considered overrated in 2026, Google has said it ignores it. Kept optionally for other systems, but robots.txt, schema and real content do the work.

Retest

Re-scoring AI visibility on the same prompt set after changes, to prove impact. "Without a retest, it's just guesses."