Glossary

The AI search vocabulary, defined neutrally.

The category still runs on competing names — GEO, AEO, ACO, AI SEO — and a fast-moving protocol layer underneath. This glossary defines every term accurately, credits the coiners, hedges what is still in pilot, and links each entry to the guide and tool that put it to work.

49 terms · Last updated June 2026


A 20 terms
ACO (Agentic Commerce Optimization)
Preparing product catalogs so AI shopping agents can evaluate, compare, and recommend them. Coined by ReFiBuy.
aco
AEO (Answer Engine Optimization)
Optimizing content to be cited in AI-generated answers — with a competing minority "Agent Engine" reading.
aeo
Agent checkout
Purchase completion by an AI agent on the shopper's behalf, protocol-mediated. In pilot as of mid-2026.
agent-checkout
Agent-readability score
Per-SKU measure of whether an AI agent can parse a PDP well enough to recommend it or draft it into a cart.
agent-readability-score
Agentic commerce
Commerce in which AI agents research products, draft carts, and — in pilot — complete checkout for shoppers.
agentic-commerce
ACP (Agentic Commerce Protocol)
The open OpenAI/Stripe protocol for agent-merchant transactions. Backs ChatGPT Instant Checkout.
agentic-commerce-protocol
Shopify Agentic Storefronts
Shopify's channel wrapping protocol integrations so merchants reach AI surfaces without custom endpoints.
agentic-storefronts
AI answer engine
A system that composes a direct, cited answer to a query instead of returning links — ChatGPT, Perplexity, AI Overviews.
ai-answer-engine
AI brand visibility
How often a brand name appears in AI answers, independent of which SKUs get resolved.
ai-brand-visibility
AI discoverability
Whether an AI engine can find and correctly identify a brand or product at all — the precondition for visibility.
ai-discoverability
AI engine optimization
Generalist near-synonym of GEO: making content and catalogs visible to AI search engines.
ai-engine-optimization
AI reputation management
Shaping how AI engines describe a brand: source corrections, review seeding, entity hygiene.
ai-reputation
AI search engine optimization
The spelled-out form of AI SEO — same umbrella, formal-document register.
ai-search-engine-optimization
AI sentiment analysis
Reading the tone and framing of AI answers that mention a brand or product.
ai-sentiment
AI SEO
The plain-English umbrella for SEO practice adapted to AI-powered search surfaces.
ai-seo
AI traffic analytics
Measuring the sessions AI engines send to a site, attributed back to prompts where possible.
ai-traffic-analytics
AI visibility
How often and how prominently an entity appears in AI-generated answers. The category's core outcome metric.
ai-visibility
AI visibility metrics
The measurement family for AI answers: citation rate, share of model, answer position, sentiment, readiness scores.
ai-visibility-metrics
AI visibility score
A 0–100 per-SKU composite of citation count, position, engine coverage, and PDP readiness.
ai-visibility-score
AP2 (Agent Payments Protocol)
The payments-layer protocol for authorizing and auditing payments made by AI agents. Pairs with UCP.
ap2
B 1 term
Brand mentions
Any reference to a brand name in an AI answer, cited or uncited — the raw signal under brand visibility.
brand-mentions
C 6 terms
ChatGPT Instant Checkout
Purchases completed inside ChatGPT, powered by ACP. Pilot, limited merchant set.
chatgpt-instant-checkout
ChatGPT Shopping
OpenAI's product-discovery experience: product results, comparisons, and draft carts inside ChatGPT.
chatgpt-shopping
Citation analysis
The systematic study of which sources an AI engine cites and why — the diagnostic layer under visibility work.
citation-analysis
Citation score
Per-SKU measure of whether AI engines actually recommend the product, sampled across buyer prompts per engine.
citation-score
Conversational search engine
A search system that handles queries as multi-turn dialogue and composes answers — ChatGPT, Perplexity, Google AI Mode.
conversational-search-engine
COSMO (Amazon)
Amazon's commonsense knowledge framework for intent-aware ranking — the system behind Rufus-era results.
cosmo
D 1 term
Draft cart
A cart assembled by an AI agent and handed to the buyer for approval. Live in ChatGPT Shopping and Buy with Pro.
draft-cart
G 2 terms
GEO (Generative Engine Optimization)
Optimizing content, products, and entities so generative AI engines surface, cite, and recommend them.
geo
Grok AI search
xAI's assistant as a search surface: live web and X retrieval plus the DeepSearch agentic research mode.
grok-ai-search
H 1 term
Hallucination detection
Catching AI engines fabricating product details — price, features, variants, availability — the PDP never claimed.
hallucination-detection
L 5 terms
LLM optimizer
The tool category for getting content or product catalogs retrieved, cited, and recommended by large language models.
llm-optimizer
LLM seeding
Placing brand and product information in the sources LLMs cite and train on, so engines mention the brand.
llm-seeding
LLM SEO
Practitioner near-synonym for GEO, emphasizing large language models as the optimization target.
llm-seo
LLM visibility
Near-synonym for AI visibility emphasizing the model as the surface where entities are cited.
llm-visibility
llms.txt
A proposed plain-text file giving LLM crawlers a curated reading list of a site's most important content.
llms-txt
M 1 term
MCP (Model Context Protocol)
The open protocol for connecting AI models to tools and data — the transport pattern in agent commerce stacks.
mcp
P 6 terms
PDP optimization
Improving product detail pages — copy, schema, metafields — so engines and agents can cite and parse them.
pdp-optimization
Perplexity Buy with Pro
Perplexity's transaction-capable shopping experience for Pro subscribers.
perplexity-buy-with-pro
Product AI visibility
The extent to which a specific product — by SKU, variant, or brand + model — is surfaced and recommended by AI engines.
product-ai-visibility
Product schema
Product JSON-LD structured data — the machine-readable identity layer AI engines extract first.
product-schema
Prompt monitoring
Watching a defined prompt set for changes in how AI engines answer — alerts on movement.
prompt-monitoring
Prompt tracking
Running a held-constant prompt set on a cadence and recording which entities get cited.
prompt-tracking
Q 1 term
Query fan-out
How AI engines expand one prompt into many retrieval queries before composing an answer.
query-fan-out
R 1 term
Rufus (Amazon)
Amazon's shopping assistant — answers product questions in the store using COSMO-era intent knowledge.
rufus
S 3 terms
Share of model
The share of an AI model's answers in a category that feature a given brand or product.
share-of-model
Share of voice (AI)
A brand's share of mentions across AI answers in a category, relative to competitors.
share-of-voice-ai
SKU-level AEO
Optimizing individual PDPs, schema, and metadata so AI engines cite specific SKUs, not just the parent brand.
sku-level-aeo
U 1 term
UCP (Universal Commerce Protocol)
Google's protocol for agent-led commerce across AI Mode, Gemini, and partner surfaces. In pilot.
universal-commerce-protocol

How to use this glossary

Each entry opens with a 2–3 sentence definition written to stand alone — the paragraph an AI engine can lift verbatim — followed by ecommerce context, a worked example, and links to the neighboring terms, the deeper guide, and the free tool that applies it. Category terms coined elsewhere are credited: ACO to ReFiBuy; the competing Agent Engine reading of AEO is noted in its entry. Terms eCommerce Insights defines — the citation score and agent-readability score — are published here as canonical definitions.

New to the category? Start with AI visibility, then GEO, then agentic commerce — the metric, the discipline, and the direction of travel.

Frequently asked questions

Why are there three names for the same discipline — GEO, AEO, and ACO?
The category has not converged on one name as of mid-2026. GEO (Generative Engine Optimization) carries the largest search volume; AEO (Answer Engine Optimization) emphasizes answer citation, with a minority "Agent Engine" reading in circulation; ACO (Agentic Commerce Optimization, coined by ReFiBuy) emphasizes catalog readiness for shopping agents. This glossary defines each neutrally — for an ecommerce catalog, the practical work overlaps heavily under all three.
Which terms should an ecommerce team learn first?
Start with the measurement pair — AI visibility and product AI visibility — then the two scores that operationalize them: the citation score and the agent-readability score. From there, GEO covers the optimization discipline and agentic commerce covers where the category is heading. Each entry links the logical next term.
Are these definitions vendor-neutral?
The category terms are defined neutrally, including terms coined by competitors — ACO is credited to ReFiBuy, and AEO's competing "Agent Engine" reading is noted in its entry. Terms eCommerce Insights itself defines (citation score, agent-readability score) are marked as such and published as canonical definitions, free to reference with attribution.
How current is this glossary?
Entries carry a last-updated stamp and hedge fast-moving claims with dates ("in pilot as of mid-2026"). Protocol entries (ACP, UCP, AP2) describe published specs rather than rumors, and engine-behavior claims state their observation basis. Entries are revised as engines and protocols change.

Vocabulary is the easy part.

See which of your products the engines actually cite — and what to fix on every PDP they skip.