AI Glossary

An autonomous software program that uses LLM-based models to perform tasks, make decisions and take actions. Can browse the web, write code and execute multi-step plans.

Content produced using AI tools (ChatGPT, Claude, Gemini, etc.). To be valuable for SEO and GEO, it must pass human editorial review and be enriched with original data and E-E-A-T signals.

Google's feature showing LLM-generated summary answers at the top of search results with cited sources. Sites selected as sources gain brand authority and referral traffic. The primary target of GEO.

The practice of optimizing content to appear as a cited source in answer engines like Perplexity, ChatGPT and You.com. Q&A format, direct answers and summarizable structure are core requirements.

A link from another website pointing to your page. Search engines treat backlinks as "votes of confidence." One of the top-3 ranking factors for over a decade. Link quality matters more than quantity.

An HTML element that tells search engines the preferred URL when similar content exists across multiple URLs. Critical for e-commerce filter pages and URL parameters. Every page should self-reference as canonical.

Google's official user experience ranking signals: LCP (loading <2.5s), INP (interaction <200ms), CLS (visual stability <0.1). One of the core pillars of technical SEO.

The crawling capacity Google allocates to your site within a given timeframe. Critical for large sites — low-quality pages waste crawl budget and prevent important pages from being indexed.

A 1-100 score predicting a domain's ranking strength based on its backlink profile. Calculated by Moz (DA) and Ahrefs (DR). Not an official Google metric, but a practical comparison benchmark.

Experience, Expertise, Authoritativeness, Trustworthiness — Google's content quality evaluation framework. Applied especially strictly for YMYL sites. AI systems also use E-E-A-T signals for source selection.

A vector representation that encodes the meaning of text in mathematical form. LLMs convert content into embedding vectors to understand semantics. Semantically rich content scores well in vector similarity.

The highlighted answer box at the top of Google search results (Position Zero) showing a direct answer from a webpage. Can be paragraph, list or table format. Strong correlation with AI Overview source selection.

Generative Engine Optimization — the practice of optimizing content to be cited as a source in AI-powered search engines. Google AI Overview, ChatGPT, Perplexity and Bing Copilot are the primary targets.

LLMs basing their responses on verifiable real-world sources rather than internal model knowledge. Original data, research citations and expert opinions increase the chance of being selected as a grounding source.

An HTML attribute telling search engines which language and region each page targets on multilingual/multi-regional sites. Incorrect implementation leads to duplicate content and wrong pages ranking in wrong countries.

Interaction to Next Paint — replaced FID in March 2024. Measures the response time for all click, touch and keyboard interactions. Target: under 200ms. Main thread blocking is the primary issue.

Google's structured database of real-world entities (people, places, organizations, concepts) and their relationships. Appearing in the Knowledge Graph provides a significant advantage for AI Overview source selection.

Largest Contentful Paint. Time for the largest image or text block to load. Under 2.5 seconds is "good." Image optimization and server response time directly impact LCP.

Large Language Model. An AI model trained on large-scale text data capable of generating human-like text. GPT-4, Claude, Gemini and Llama are examples. The primary target systems for GEO strategy.

A standard file that tells AI systems how a website wants its content used — robots.txt for AI. One of the core technical components of a GEO strategy.

The practice of crafting optimal input instructions to get the best responses from LLM-based AI systems. In the SEO context, relevant for understanding how AI systems interpret and use content.

Retrieval-Augmented Generation. The architecture where LLMs retrieve relevant documents from the web before generating an answer. Content must pass accessibility, relevance and trust filters to become a source.

Structured data (JSON-LD) providing machine-readable context about page content to search engines. Enables rich results and increases AI Overview source selection probability. Critical technical signal for GEO.

Search Engine Results Page. The page Google displays for a user query. 2025 SERPs include AI Overview, featured snippets, local pack, shopping results and knowledge panels alongside organic listings.

When a site comprehensively covers a topic and is recognized by search engines as an expert source. The pillar-cluster model is the core methodology. Increases AI citation probability.

The experience where users find answers on the results page without clicking any link. Growing with AI Overview. Brands appearing as AI sources gain brand visibility even without click traffic.