--- title: "Glossary" url: https://www.thinkai.it/en/glossary/ date: 2026-10-13 modified: 2026-10-13 author: "Admin" type: page lang: it --- # Glossary Every term we use on this site, explained in one sentence, with the everyday word it corresponds to and the page where it is used. Six AI platforms · One method · Verifiable numbers ## Concepts The words that describe the problem. GEO is explained in full in the article [what Generative Engine Optimization is](https://www.thinkai.it/en/blog/what-is-geo-generative-engine-optimization/); GEM on the [Generative Experience Marketing](https://www.thinkai.it/en/gem-generative-experience-marketing/) page. ### AI Brand Governance The work of measuring and steering how generative AI systems describe, position, and recommend a brand. Keeping what AI says about a company under control ### AI visibility How much and how a brand appears in AI answers: whether it is cited, in what position, in what tone. Being present in AI answers ### GEO, Generative Engine Optimization The set of interventions on sources, structure, and clarity of information so that AI models understand and cite a brand. SEO for AI answers ### GEM, Generative Experience Marketing The discipline that coordinates a brand’s organic presence in AI answers (GEO) with its paid presence in the same spaces, such as ChatGPT Ads and Google’s ads in AI Overviews and AI Mode. Making the AI answer and the ad next to it say the same thing ### AEO, Answer Engine Optimization Preparing content so that it answers questions directly, in the format that answer engines and AI pick up. Writing answers, not just pages ### AI Overview The AI-generated summary Google shows at the top of search results, with the sources it was built from. Google’s answer before the links ### LLM, Large Language Model The language model behind ChatGPT, Gemini, Claude, and the other assistants: it generates text from what it read in training and what it retrieves at query time. The engine of the AI assistant ### AI Perception Gap™ The distance between how a company wants to be perceived and how AI systems actually describe it. What AI gets wrong about you ## Method and tools How we work. The four phases are described on the [ThinkAI BrandOS™](https://www.thinkai.it/en/method/) page; the starting point is [the AI Brand Audit](https://www.thinkai.it/en/ai-brand-audit/). ### ThinkAI BrandOS™ ThinkAI’s method in four repeating phases: Monitor, Diagnose, Orient, Correct. Our way of working, in cycles ### AI Brand Audit The initial measurement: 80 questions across 6 AI platforms, recorded and traced back to their sources, with a competitor comparison and an action plan. The starting snapshot ### Light Audit The free, short version of the audit: 16 questions on ChatGPT, Gemini, and Perplexity, results presented in a 20-minute call with a written summary. A taste of the audit ### IntentIQ ThinkAI’s platform that asks several AIs the same questions, records the answers and sources, and tracks how they change over time. The tool that asks the questions and keeps score ## The documents the client receives Each phase produces a document. They are delivered with the audit and at every governance cycle; [IntentIQ](https://www.thinkai.it/en/intentiq/) feeds them with the recorded answers. ### Brand Health AI Dashboard™ The dashboard that gathers the indicators of how AI describes the brand, updated every cycle. The control panel ### AI Perception Gap Report™ The document that quantifies the distance between the intended positioning and the one perceived by AI, with the risk areas. The report on what AI gets wrong ### Source Intelligence Map™ The list of sources AI uses to talk about a brand, with the weight of each. Where AI got what it says ### RIM Matrix™ The matrix that ranks interventions with the formula Impact² / Effort: first those that change the brand’s representation the most with the least work. The priority list ### Content Architecture Plan™ The plan of content and structured data to build or fix so that AI finds clear, complete information. What to write and where ### Governance Cycle Report™ The end-of-cycle report: interventions made, AI answers before and after, next priorities. The minutes of each cycle ### GEM Readiness Report How the brand appears in both layers of an AI answer, organic and paid, question by question, with sources and competitors’ ads. The snapshot of answer and ads together ### Paid Layer Plan The plan for the paid layer: where to buy, on which questions, and with which message, with priorities and expected results. The ads plan ## Indicators The numbers we use to measure results, cycle after cycle. They are also described in the metrics section of the [method](https://www.thinkai.it/en/method/). ### AI Perception Score™ A 0 to 100 index of consistency between how the brand wants to be perceived and how AI describes it. The overall grade ### AI Presence Rate™ Percentage of relevant questions in which the brand is cited correctly in AI answers. How often you show up ### Citation Quality Index™ Accuracy, tone, and competitive framing with which AI describes the brand when it cites it. How they talk about you ### Silent Warning Frequency™ How many times per cycle AI answers contain an implicit warning about the brand, such as “some users report issues.” AI’s little digs ### AI Readability Score™ How well the brand’s content is correctly understood by AI systems. How well AI understands you ### Correction Effectiveness Rate™ Percentage of interventions, organic and paid, that produce a positive change in the next cycle’s answers. How many interventions work ### Answer Consistency Score™ Consistency, from 0 to 100, between the brand’s organic description in the AI answer and the paid ad on the same question. Do the answer and the ad agree? ## Want to see these numbers for your brand? The free Light Audit calculates them on 10 questions and 2 AI platforms, in five business days.