AI visibility glossary

Whether AI answer engines mention, recommend, omit, or misrepresent your brand for buyer-intent prompts. It's measured separately from search rankings because the two can diverge.

The share of tested prompts where an engine names and recommends you (not merely mentions you), reported as a numerator over a denominator — e.g. 12 / 40 prompts.

The share of tested prompts where your brand appears in the AI answer at all, in any form (mentioned, cited, or recommended). A broader signal than recommendation share.

The share of tested prompts where an engine cites a page you own as a source. Most relevant on citation-heavy engines like Perplexity that footnote their answers.

The share of tested prompts where your brand is absent from the answer entirely — neither mentioned nor cited. Often the most actionable number in a first audit.

When an engine names you but states something inaccurate or stale — wrong pricing, outdated features, a competitor's framing. Captured as evidence and traced to a likely source.

Generative Engine Optimization — improving the public evidence (content, sources, structured data) so AI answer engines can accurately include and describe you. A layer on top of solid SEO, not a replacement.

Answer Engine Optimization — a near-synonym for GEO that emphasizes making content directly answer the questions engines synthesize answers from. Often used interchangeably with GEO.

Search Engine Optimization — optimizing for position in a ranked list of links. Still essential: it produces the crawlable, accurate, useful pages that AI answers are built from.

A prompt a real buyer would ask while evaluating tools — "best X for Y", "X vs Z", "alternatives to W". Audits test a fixed set mapped to persona and funnel stage.

A third-party source (review site, listicle, comparison page) that drives citations in your category but omits or misrepresents you. The map of source gaps is where most fixes start.

A system that responds to a query with a synthesized answer rather than a list of links — ChatGPT, Google AI Overviews / AI Mode, Perplexity, Gemini, Claude.

Retrieval-Augmented Generation — when a model retrieves live documents and uses them to ground its answer, rather than relying only on what it learned in training. It's why your crawlable pages can be cited in real time.

Tying a model's answer to specific retrieved sources so claims trace back to real documents. Grounded answers are more likely to cite you accurately; ungrounded answers lean on trained memory.

Standard machine-readable markup (schema.org) that states your facts unambiguously — name, category, pricing, FAQs. There's no special "AI markup"; accurate standard schema is the lever.