Glossary

Action · One recommended piece of work in the Action Centre, scored for value and grounded in run data. Agent Experience · The tab of the GEO Audit with two site-wide checks: whether the crawlers AI engines use for their answers can reach your site, and whether your homepage carries the block that names your business. There is no Agent Experience score. Agency Admin · The console where an agency manages clients, prospects, team, client logins, its shared taxonomy and billing. AI answer crawler · A crawler an AI engine sends to read a page before citing it. Blocking one in robots.txt keeps you out of that engine’s answers; training crawlers are different, and blocking them is a choice. Background searches · The sub-queries an AI platform ran while composing an answer, where the platform exposes them. Also called fan-out queries. Board · The Trello-style view in the Action Centre for tracking committed actions through Backlog, In Progress and Done. Branded prompt · A prompt that contains your brand name. Measures how AI describes you, not whether it finds you. Excluded from rolled-up metrics and checked weekly. Citation · An answer linking to, or listing as a source, a page on a domain. Citation Share · Your domain’s share of all citations across tracked brand domains. Client workspace · In an agency, a client’s own organisation with its own sites, logins and reports. Connector key · The organisation-scoped key an owner creates under Integrations to feed Looker Studio. Shown once and revocable at any time. Content pack · The LinkedIn post, thread, FAQ block, video script and JSON-LD generated from a finished draft. CRAFT · The five checks on the GEO Audit: Crawlability, Recognition, Answers, Format and Trust. Data as of · The timestamp of the run a page is reading from. Discovery scan · The on-demand scan behind the AI Search Index that finds which of your pages sit in the shared source pool. Draft · A piece written by the content agent in the Content Studio. Evidence · The live checks on a draft showing how many retrieval conditions it meets. Gap · A prompt where competitors are mentioned or cited and you are not. GEO · Generative Engine Optimization: writing and structuring pages so AI platforms can extract, quote and cite them. Kevin · Koalr’s assistant, in the bottom corner of every page (right by default). Location fragment · The bracketed part of a shared prompt, such as [in {location}]. A regional client is asked the prompt with their town in place; a nationwide client is asked it without the fragment. Viewer, Editor, Manager · The three per-site access levels for members. Owners bypass them. MCP · Model Context Protocol. Koalr’s MCP server lets AI agents read your visibility data. Mention · A brand named in an answer’s text. Multi-Site Dashboard · Every site in the organisation compared side by side. Called Agency Dashboard for agencies. Not optimised · A page you have deliberately excluded from the GEO Audit, Actions and Kevin’s advice. Opportunity · A content action in the Content Studio: rival page, own winner, gap or gap cluster. Organisation · Your Koalr account’s workspace: sites, members, billing. Own winner · One of your pages already cited on several prompts, and the companion piece it suggests. Page to beat · The rival page (or your own working page) a draft is written to out-cover. Pooled prompts · The agency prompt allowance, contributed per site and spent across every client. Prompt · A real question Koalr asks AI platforms on your behalf. Prompt Visibility Rate · Of the AI answers to your tracked prompts that name a brand, the share that name yours. Answers that name no brand are left out, and a link to your site without your name is not a mention. The headline metric. Prospect · A free pitch workspace an agency uses to show a prospective client their data: one analysis, never refreshed. Refresh workbench · The page read through the six ANSWER letters, with the additions that would help AI quote it. Nothing already on the page is rewritten. Revision · A refreshed draft of an existing page: additions only, with everything already there kept exactly as written. Run · One pass over every active prompt on every tracked platform, stored in full. Sentiment · How positively AI describes you when it mentions you, 0 to 100. Share of Voice · Of all brand mentions across your answers, the slice that is yours. Shared taxonomy · An agency’s set of topics and prompts, locked on every client site it is applied to, so clients compare like for like. Site · One brand in one country: domain, competitors, topics, prompts and history. The unit of tracking and billing. Site GEO Score · The average GEO score of your scored pages, out of 100, from the GEO Audit. Source pool · The set of pages an AI platform can pick from when composing an answer to a prompt. Tone of voice document · Your saved description of how you write, with real sample sentences, which the content agent writes to. Topic · A group of related prompts. What next · The guided checklist and weekly feed opened from the Dashboard.