Glossary

AI search & discovery, defined

The vocabulary of GEO, AEO, and AI-powered product discovery — in plain language, for ecommerce teams.

AI answer engine: An AI answer engine is a search system that responds to a question with a synthesized, conversational answer rather than a ranked list of links. ChatGPT, Perplexity, Google AI Overviews, and Claude are answer engines: they recommend specific products and brands directly, citing sources they can read and trust.
AI citation share (visibility share): AI citation share is the portion of AI-generated answers in a category that mention or cite a given brand — the answer-engine analogue of search ranking. If ChatGPT recommends dresses to a thousand shoppers and your brand appears in a hundred of those answers, your citation share is 10%.
AI Overviews: AI Overviews are Google's AI-generated answers that appear above the classic organic results, synthesizing a response from sources Google's models can read and cite. For many shopping queries, the Overview — not the first blue link — is now the first thing a shopper sees.
AI product discovery: AI product discovery is on-site search and browsing powered by machine understanding of products and shoppers: conversational search, style- and fit-aware recommendations, and complete-the-look suggestions in place of keyword search and static category grids.
AI shopping agent: An AI shopping agent is an AI assistant that acts on a shopper's behalf — researching options, comparing products, filling carts, and increasingly completing checkout. Agentic shopping shifts the buying decision to software that only recommends products it can read, verify, and transact against.
AI visibility: AI visibility is how findable and recommendable a brand is across AI surfaces — answer engines, AI Overviews, and shopping agents. It spans access (can AI systems read you), understanding (do they interpret your products correctly), and citation (do they actually recommend you).
Answer Engine Optimization (AEO): Answer Engine Optimization (AEO) is the practice of making content and products understandable and citable by AI answer engines — systems like ChatGPT, Perplexity, and Google AI Overviews that reply with a synthesized answer instead of a list of links. AEO and GEO describe the same discipline.
Fit-aware recommendations: Fit-aware recommendations match products to a shopper's physical traits and fit preferences — proportions, rise, drape, cut — inferred from behavior and product understanding rather than size quizzes or measurements. The goal: fewer wrong-fit purchases, fewer returns, higher conversion.
Generative Engine Optimization (GEO): Generative Engine Optimization (GEO) is the practice of structuring a brand's products and content so generative AI engines like ChatGPT, Google AI Overviews, and Perplexity can understand, recommend, and cite them. Where SEO competes for a ranked link, GEO competes to be the answer itself.
Generative Search Optimization (GSO): Generative Search Optimization (GSO) is a synonym for Generative Engine Optimization (GEO): optimizing content and product data so generative AI search systems can understand, recommend, and cite it. Different teams use GSO, GEO, or AEO — the practice is the same.
LLM crawler: An LLM crawler is a bot that fetches web pages on behalf of an AI system — for training data, for live answer retrieval, or for a user's specific question. GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, and Google-Extended are LLM crawlers; blocking them makes a store invisible to the engines they feed.
llms.txt: llms.txt is a plain-markdown file at a site's root that gives AI assistants a structured summary of the site: what the company is, what matters, and where the substance lives. Like robots.txt for permissions or sitemap.xml for URLs, llms.txt is the entry point an LLM reads first.
Product feed enrichment: Product feed enrichment is the process of adding structured attributes to a product catalog beyond what merchants enter by hand — fit, silhouette, fabric behavior, occasion, style descriptors — so search engines, AI answer engines, and recommendation systems have real data to reason about.
Retrieval-augmented generation (RAG): Retrieval-augmented generation (RAG) is the technique AI systems use to ground answers in live sources: retrieve relevant documents first, then generate the answer from them, citing what was retrieved. It's why being retrievable — crawlable, structured, quotable — determines whether a brand gets cited.
Semantic product data: Semantic product data describes products in terms of meaning rather than keywords: fit, silhouette, material behavior, occasion, styling intent, and the relationships between them. It's what lets an AI system understand that a "relaxed linen blazer" suits a summer office and a beach wedding alike.
Structured data (schema.org): Structured data is machine-readable markup — usually JSON-LD using schema.org vocabulary — embedded in web pages to state facts explicitly: this is a Product, it costs this much, it has these reviews, this Organization sells it. Search engines and AI systems read it as ground truth about a page.