The AI Search Glossary

Plain-English definitions for GEO, LLM SEO, AI visibility, prompt tracking and the rest of the vocabulary you'll need in the AI-first era of search.

GEO (Generative Engine Optimization)

Optimizing your brand to be recommended inside AI-generated answers.

GEO is the practice of optimizing brand entities, content, and citations so that LLMs like ChatGPT, Gemini and Claude mention your brand when buyers ask category-defining questions. It is the AI-era successor to SEO.

LLM SEO

SEO tactics adapted for Large Language Models instead of traditional search engines.

LLM SEO focuses on getting cited and surfaced inside AI answers rather than ranking blue links. Tactics include entity clarity, comparison content, community presence, and structured data.

AI Visibility

How often your brand appears inside AI-generated answers across buyer-intent prompts.

AI Visibility measures the presence of your brand inside answers from ChatGPT, Gemini, Claude and Perplexity for a defined prompt set. It's the AI equivalent of impression share.

AI Share of Voice (AI SoV)

Your brand's share of total brand mentions across tracked AI answers.

AI SoV (%) = (your brand mentions ÷ total brand mentions across all tracked prompts) × 100. The cleanest metric for benchmarking AI visibility against competitors.

Prompt Tracking

Systematically running buyer-intent prompts across AI models to monitor brand mentions.

Prompt tracking turns ad-hoc 'I asked ChatGPT once' checks into a repeatable measurement system. Hundreds of prompts × multiple models × weekly cadence = a real signal.

AI Prompt Tracker

Software that runs buyer-intent prompts across LLMs and reports brand mentions.

An AI Prompt Tracker (like ClickAI) automates running prompts across ChatGPT, Gemini and Claude, scores brand and competitor mentions, and exports the results to Excel and PDF.

Buyer-Intent Prompt

A prompt a real prospect would type when they are close to a purchase decision.

Examples: 'best CRM for early-stage startups', 'alternatives to HubSpot under $50/mo', 'is Notion good for engineering teams'. These are the prompts worth tracking.

Brand Mention

An occurrence of your brand name inside an AI-generated answer.

Detected via whole-keyword, case-insensitive matching across the full AI response. Mentions are counted per prompt and rolled up into Share of Voice.

Citation

A source the LLM links to or quotes when generating an answer.

Modern LLMs increasingly cite sources (Reddit, G2, YouTube, news sites). Earning a citation on a high-authority source is one of the strongest GEO levers.

Grounding

When an LLM bases its answer on retrieved real-time or trusted sources instead of training data.

Grounding (often via RAG or web search) improves answer accuracy and increases the importance of being present on authoritative sources.

Hallucination

When an LLM confidently states something that isn't true.

For brand monitoring, hallucinations matter when LLMs mis-attribute features or pricing to your brand. Prompt tracking surfaces these so you can correct them.

Entity

A real-world thing (brand, product, person) the LLM recognizes as a distinct concept.

Entity clarity — clean Wikidata, consistent sameAs links, structured Organization schema — directly improves how often an LLM recommends you for your category.

Share of Answer

Percentage of tracked AI answers that mention your brand at least once.

Different from SoV: SoA counts answers, not mentions. Useful when one answer mentions a brand many times and would otherwise skew the score.

Multi-Model Tracking

Running the same prompt across ChatGPT, Gemini, Claude and Perplexity.

Single-model tracking is misleading — each model answers differently. Multi-model coverage is required for a true AI visibility picture.

AI SERP

The synthesized answer panel an AI assistant returns instead of a list of links.

The AI SERP is the new battleground. Position #1 in Google matters less when the AI summarizes the answer before the user ever clicks.