Glossary.
The terms used in Visiqn reports and on this site, defined plainly.
- AI assistant
- A conversational AI product that answers questions in natural language — ChatGPT, Gemini, Claude and Perplexity are the four Visiqn measures.
- AI visibility
- Whether, and how, AI assistants name your business when a buyer asks for a recommendation. See the guide.
- AI-readiness score
- A 0–100 score for a website based on fifteen checks across structured data, entity clarity, citation-worthiness, freshness and crawlability.
- Buyer-intent question
- A question a person asks when they are close to choosing a business: recommendations, comparisons, prices, trust, urgency. Visiqn generates thirty per report.
- Citation-worthiness
- Whether a page contains a specific, clear statement an assistant can quote as an answer.
- Clean session
- A query run with no prior conversation or personalisation, so the answer reflects what the assistant believes in general rather than what it has learned about the person asking.
- Competitor ranking
- The list of businesses assistants named for your questions, ranked by how often each appeared.
- Crawlability
- Whether a retrieval system can reach and read your pages: robots rules, sitemaps, and no walls in front of content.
- Entity clarity
- How certain a machine can be about which business you are — one name, one address, one description, consistent everywhere. See the guide.
- Factual error
- A claim an assistant makes about your business that is wrong: hours, price, service, address, description.
- Fix list
- The prioritised list of changes at the end of every report, ordered by expected impact on mention rate.
- Freshness
- Whether a site shows signs of being current: visible dates, recent updates, no abandoned pages.
- Market
- One city, region or country in which a report measures a business. Multi-location businesses are measured per market.
- Mention rate
- The percentage of the 120 answers in a report that named your business.
- Model version
- The specific version of an assistant's underlying model at the time of the query, recorded in the report appendix.
- Retrieval
- The process by which an assistant fetches web pages to answer a question, as opposed to answering from memory alone.
- Schema / structured data
- Machine-readable markup (usually JSON-LD) that states facts about a business — name, address, hours, services — in a standard format.
- Source tracing
- Linking each claim an assistant makes back to the page it cited or matched, so the claim can be corrected at its origin.
- White-label
- A report delivered under an agency's own branding, with Visiqn unnamed.