> ## Documentation Index > Fetch the complete documentation index at: https://docs.perquery.ai/llms.txt > Use this file to discover all available pages before exploring further. # Concepts > The main objects of Perquery (sites, prompts, personas, competitors, scans) and how Perquery calculates its metrics. This page explains the objects that you work with in Perquery and the metrics that Perquery calculates. For the product label of each term in English and Turkish, see the [Glossary](https://docs.perquery.ai/en/glossary.md). ## Organization An organization is your account in Perquery. It has the members, the roles, the plan and the sites. Each member has one or more roles. A role controls what the member can see and do. See [Organization and team](https://docs.perquery.ai/en/account/organization-and-team.md). ## Site A site is one brand that Perquery tracks. In the product, a site is a **Website**. A site has: * A **Brand name**, which answers and reports refer to. * A domain, for example `flowdesk.io`. You cannot change the domain after you create the site. * An optional **Category**, for example "B2B SaaS, CRM". * A market: a country and the language that your buyers ask in. Perquery generates prompts, competitors and answers for this market. * **Brand aliases**: other names of your brand. Perquery uses them to find your brand in AI answers. All data in Perquery belongs to one site. Use the site switcher in the sidebar to change the site. The number of sites depends on your plan. See [Add a site](https://docs.perquery.ai/en/sites/add-a-site.md). ## Prompt A prompt is a question that a buyer asks an AI engine, for example "Best CRM tools for agencies". Perquery sends each tracked prompt to each selected AI engine and reads the answers. * When you add a prompt, a scan starts for it. * You can change the text of a prompt only until Perquery measures it. * A prompt can have a target page: the page of your site that must answer the prompt. The number of prompts depends on your plan. See [Prompts](https://docs.perquery.ai/en/monitor/prompts.md). ## Tracking group A tracking group is a label that you give to prompts, for example a product line or a topic. A prompt can be in more than one group. **Group Performance** shows the average metrics of the prompts in each group. See [Prompt groups](https://docs.perquery.ai/en/monitor/prompt-groups.md). ## Persona A persona is a type of buyer. Each persona is a system prompt: a description that tells the AI engine who asks the question. Perquery asks each prompt as each persona and measures each persona separately. * Each site has one **Baseline**. The baseline is **Raw query** (no persona) or one persona that you describe. Charts open on the baseline. Your plan does not count the baseline. * You cannot change the baseline after the first scan. This keeps its trend comparable. * A persona has a **Persona kind**: **Demographic**, **Need-based**, **Behavioural** or **Custom**. * You can mark a persona as a **Target**: a segment that your brand is positioned for. * A new or changed persona collects data from the next scan. It has no historical data. See [Personas](https://docs.perquery.ai/en/monitor/personas.md). ## Competitor A competitor is a brand that you compare your brand with. Perquery measures tracked competitors in every scan and shows them in every comparison. Answers can also name brands that you do not track. Perquery counts these brands too. **Competitor Analysis** groups them as **Other**. When you remove a competitor, past scans keep their numbers. See [Competitors](https://docs.perquery.ai/en/monitor/competitors.md). ## AI engine An AI engine is an AI assistant or AI search product that Perquery asks, for example ChatGPT or Gemini. You select the AI engines for each site. See [AI engines](https://docs.perquery.ai/en/ai-engines.md). ## Scan A scan is one run of measurement. In the product, it is a **Visibility scan**. A scan asks the AI engines your prompts, for the baseline and for each persona, and records each answer. * Scans run on the schedule of the site. You set the schedule in the site settings. * Depending on your plan, a scan can ask a share of the prompts. Each prompt is then asked once in a period that your plan sets. * A scan has a status: **Running**, **Finalising**, **Completed**, **Partly completed**, **Failed** or **Cancelled**. See [Scans and schedule](https://docs.perquery.ai/en/sites/scans-and-schedule.md). ## How Perquery reads an answer For each answer, Perquery finds out: * If the answer shows your brand. The brand is in the answer when the text names it or when the answer cites a page on your domain. * How the answer shows your brand. A mention: the text names the brand. A citation: the answer links to your domain as a source. A recommendation: the answer recommends the brand. * The position of your brand: its place in the list of the answer, or the order in which the answer first names the brands. * The tone in which the answer talks about your brand. * Which tracked competitors and other brands the answer names. ## Metrics | Metric | What it shows | | - | - | | **Visibility** | How often and how high your brand appears in the answers, from 0 to 100. An answer without your brand counts as 0. A first place counts more than a lower place. | | **Share of Voice** | Your share of all brand mentions in the answers, as a percentage. All named brands count, also the brands that you do not track. | | **Avg. Rank** | Your average position in the answers that name your brand. Lower is better: 1 is the first brand named. | | **Avg. Sentiment** | The average tone of the answers that show your brand. The product shows it as **Positive**, **Neutral** or **Negative**. | | **Domain Rating** | The backlink rating of your domain, from 0 to 100, at each scan. | All metrics follow the filters at the top of the page: the persona, the date range and the AI engine. When you select only some AI engines, the numbers cover only those engines. Do not compare them with numbers for all engines. ## Opportunity An opportunity is a change that can make your brand more visible. Each opportunity has a type (**Technical**, **Content** or **Off-page**) and an **Opportunity score**. A higher score means a higher priority. See [Opportunities](https://docs.perquery.ai/en/act/opportunities.md). This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.