Glossary

Every term in GEO Radar in plain words, with the guide that explains it.

Every term you meet in GEO Radar, in plain words, with the guide that explains it in full.

Prompts and answers

Prompt: A question buyers ask AI, tracked every night. See Prompts.
Topic: A group of related prompts, such as a product line or use case.
Tag: An optional label across topics, used to filter reports.
AI model: An assistant GEO Radar asks: ChatGPT, Perplexity, Google AI Overview, Google AI Mode, Gemini or Copilot. Each project uses three.
Scan: The nightly run that asks every enabled prompt on the project’s three models. See Scans and data freshness.
Answer: One model’s reply to one prompt. Also called a chat.
Answer that suggests tools: An answer that recommends at least one product. The base for Recommended.
Not read yet: Answers that have arrived but haven’t been read for roles, ranks and sentiment. Shaded on charts.

Tracked brand: Your brand or a competitor you added in Brands.
Mention: An answer names a tracked brand.
Role: How an answer treats a brand: recommended, compared, mentioned or warned against.
Visibility: Share of answers that name the brand.
Recommended: Share of answers that suggest tools which recommend the brand.
Position / rank: Where the brand appears in the answer; #1 is named first.
Sentiment: How positively the answer talks about the brand, 0–100, where 50 is neutral.
Share of voice: The brand’s share of all mentions, with earlier mentions weighing more.
Previous period: The period of the same length just before the selected range. Every change compares with it.
—: Not enough answers for a fair number. Never the same as zero. See Metrics explained.
AI also names: Brands AI mentions that you don’t track yet.

Source: A web page an answer used. Grouped by page (URL) or by website (domain).
Cited: Shown to the reader as a link or footnote.
Retrieved / reported only: Read by the model but not shown in the answer.
Appearance: One source in one answer.
Citation rate: Cited appearances divided by all appearances.
Answer reach: Share of all answers that use a source.
Competitor gap: A cited source in an answer that names a competitor and not you. See Gap analysis.
Brand omission: A source in an answer that doesn’t name you.
Opportunity: How much a source is worth winning, compared with the other sources in view.
Domain type / URL type: The kind of site (Community, Review, Editorial…) or page (Article, Comparison, Video…).

Recommendation: An action to improve visibility, suggested by GEO Radar or saved by your team. See Recommendations.
Suggested: Proposed from your data and not yet accepted.
Verified fixed: A done action whose problem GEO Radar no longer detects.
Before / after window: The 14 days before and after an action was completed, compared on Impact.
Measuring: An action whose after window is still running.
Used as a source: Share of answers that cite your own domain.
robots.txt: The file that tells crawlers which pages they may read. See Crawlability.
Search / training / user-request bot: Crawlers that build AI search indexes, collect training data, or fetch a page live for a person.
Verified crawler: A visit from the company’s published IP addresses. See Crawl insights.
AI referral: A person who clicked a link to your site inside an AI assistant.

Workspace: Your team’s space: projects, members and prompt capacity.
Project: One brand and its website, with its own prompts, models and reports.
Prompt capacity / allocation: How many prompts the workspace may scan, and how many each project gets. See Projects.
Ready / Draft: Whether a project is scanned every night, or waiting for an allocation.
Owner, admin, editor, viewer: The four roles. See Members and roles.