The AI Commerce Glossary
UCP, ACP, MCP, AEO, GTIN, BuyAction — AI commerce has too many acronyms. Here is every term you need to know, explained in plain English. Bookmark this page and come back whenever you hit a term you don't recognize.
A
A2A (Agent-to-Agent Protocol)
Google's protocol for AI agents to communicate with each other.
A2A is an open protocol developed by Google that enables AI agents to discover, authenticate, and collaborate with one another. It powers multi-agent workflows where a primary AI agent can delegate specialized tasks — such as inventory lookups or shipping estimates — to other agents. A2A complements MCP and UCP by focusing on inter-agent communication rather than tool access or commerce transactions.
ACP (Agentic Commerce Protocol)
OpenAI and Stripe's protocol enabling AI agents to complete purchases.
ACP is the commerce protocol jointly developed by OpenAI and Stripe that allows AI agents to handle the full purchase lifecycle on behalf of consumers. It powers ChatGPT's Instant Checkout feature, enabling one-click purchases without leaving the conversation. ACP transactions currently carry a 4% commission fee charged to merchants through Stripe.
AEO (Answer Engine Optimization)
Optimizing content to appear in AI-generated answers rather than traditional search results.
AEO is the practice of structuring your content so that AI-powered search engines cite it when generating answers. Unlike traditional SEO, which optimizes for link clicks on a results page, AEO focuses on being the authoritative source an AI model references. This involves clear, factual writing, strong structured data, and content that directly addresses user intent. AEO is becoming essential as zero-click searches grow.
Agent Ready Score
AgentReadyHQ's proprietary metric measuring AI discoverability and transactability.
The Agent Ready Score is a 0-100 metric that evaluates how well an e-commerce store is optimized for AI shopping agents. It analyzes over 20 factors across six categories: structured data quality, product identifier coverage, protocol readiness (UCP/ACP), feed optimization, policy accessibility, and technical SEO. Higher scores correlate with greater visibility in AI shopping recommendations from platforms like Google AI Mode, ChatGPT, and Perplexity.
Agentic Commerce
AI agents autonomously discovering, comparing, and purchasing products for consumers.
Agentic commerce describes the emerging paradigm where AI agents act as autonomous shopping intermediaries. Rather than a consumer manually browsing, searching, and checking out, an AI agent handles the entire journey from product discovery through purchase completion. This shift is being driven by protocols like UCP and ACP, and platforms such as Google AI Mode and ChatGPT Shopping. Merchants must adapt their stores to be machine-readable and transactable to participate.
AI Mode (Google)
Google's AI-powered search experience with native product recommendations and checkout.
Google AI Mode is the conversational AI layer within Google Search that generates synthesized answers, product comparisons, and shopping recommendations. Unlike traditional Google Shopping, AI Mode can present products contextually within a conversation and enable native checkout. Google currently charges 0% commission on AI Mode transactions, making it an attractive channel for early-adopting merchants.
AI Shopping Agent
An AI system that can browse, compare, and purchase products autonomously.
AI shopping agents are software systems powered by large language models that can perform end-to-end shopping tasks on behalf of consumers. They interpret natural-language requests, search across multiple stores, compare products on price and features, and complete transactions. Major examples include ChatGPT Shopping, Google AI Mode, and Perplexity Shopping. For merchants, visibility to these agents depends on structured data, product identifiers, and protocol implementation.
B
BuyAction Schema
Schema.org markup that tells AI agents a product can be purchased directly.
BuyAction is a schema.org action type that explicitly signals to AI agents and search engines that a product is available for immediate purchase at a given URL. It goes beyond standard Product schema by encoding the transactional intent and purchase endpoint. Only an estimated 10 to 100 domains globally currently implement BuyAction, making it a significant competitive differentiator for early adopters in AI commerce.
C
ChatGPT Shopping
OpenAI's shopping feature that lets ChatGPT recommend and sell products.
ChatGPT Shopping is OpenAI's integrated commerce experience that allows users to discover, compare, and purchase products directly within the ChatGPT interface. It uses the ACP protocol with Stripe for payment processing and charges a 4% Instant Checkout commission. Products appear as rich cards with images, prices, and one-click buy buttons. Merchants need optimized product feeds and Stripe integration to participate.
Conversion Rate (AI-Referred)
The percentage of AI-referred visitors who make a purchase.
AI-referred conversion rate measures how effectively traffic from AI shopping agents translates into sales. Industry data from Similarweb shows an average AI-referred conversion rate of 11.4%, compared to 5.3% for traditional organic search traffic. This higher rate reflects the fact that AI agents pre-qualify buyer intent before sending users to a store, resulting in more purchase-ready visitors.
Crawlability (AI)
How easily AI agents can access and understand your store's product data.
AI crawlability measures whether AI shopping agents can effectively discover and parse your product catalog. Unlike traditional SEO crawlability, which focuses on HTML accessibility for search engine spiders, AI crawlability requires structured data (JSON-LD schema), machine-readable product identifiers (GTINs), clean API endpoints, and proper robots.txt configuration that does not block AI agent user agents like GPTBot and ClaudeBot.
D
Data Feed
A structured file containing your product catalog submitted to shopping platforms.
A data feed (also called a product feed) is an XML, CSV, or JSON file that contains structured information about every product in your catalog, including titles, descriptions, prices, availability, images, and identifiers. Data feeds are submitted to platforms like Google Merchant Center, Facebook Commerce, and increasingly to AI shopping services. Feed quality directly impacts how accurately AI agents represent and recommend your products.
Discoverability (AI)
The likelihood that AI agents will find and recommend your products.
AI discoverability represents how probable it is that an AI shopping agent will surface your products when a consumer makes a relevant request. It is determined by multiple factors: the completeness and accuracy of your schema markup, whether you have valid product identifiers (GTINs), the quality of your data feeds, your protocol implementation status (UCP/ACP), and whether your robots.txt allows AI crawler access. Low discoverability means your products are invisible to the AI commerce ecosystem.
E
EAN (European Article Number)
A 13-digit product identifier used primarily in Europe, part of the GTIN family.
EAN is a standardized barcode and product identification system used predominantly in European and international markets. The 13-digit EAN is functionally equivalent to a UPC with a leading zero and is part of the broader GTIN (Global Trade Item Number) standard. AI shopping platforms require GTINs (including EANs) to match products across merchants, verify authenticity, and provide accurate comparison shopping results.
F
Feed Management
Creating, optimizing, and distributing product feeds to multiple shopping platforms.
Feed management is the operational process of generating product data feeds, transforming them to meet each platform's specifications, and distributing them to destinations like Google Merchant Center, Meta Commerce, and AI shopping services. Specialized tools such as DataFeedWatch, Feedonomics, and GoDataFeed automate this process. Effective feed management ensures your products appear with accurate, complete information across all AI and traditional shopping channels.
First-Party Data
Product and customer data you own directly, increasingly valuable in AI commerce.
First-party data refers to information collected directly by your business, including product specifications, pricing, inventory levels, and customer interactions. As AI agents prioritize authoritative sources, first-party data becomes a strategic asset. Stores that publish comprehensive, accurate product information directly on their own domains give AI agents a trustworthy source to reference, rather than relying on third-party aggregators that may have outdated or incomplete data.
G
GEO (Generative Engine Optimization)
Optimizing content specifically for AI-powered generative search engines.
GEO focuses on making your content the source that generative AI engines cite when producing answers. While AEO is broader and covers any answer-engine format, GEO specifically targets generative models like Google Gemini, ChatGPT, and Perplexity that synthesize new text from multiple sources. GEO strategies include providing unique data points, using clear factual statements, and implementing comprehensive structured data that models can easily extract and reference.
Google Merchant Center
Google's platform for managing product data feeds, required for AI Mode listings.
Google Merchant Center is the central hub where merchants upload and manage product data for Google Shopping, Google AI Mode, and other Google commerce surfaces. It validates feed quality, monitors product status, and serves as the data backbone for AI-powered shopping recommendations. Major changes are coming in 2026: a product ID format split in March and the legacy Content API sunsetting in August. Merchants should prepare for both transitions to maintain uninterrupted AI visibility.
GTIN (Global Trade Item Number)
The universal product identifier system. 87% of stores lack GTINs, the #1 audit issue.
GTIN is the umbrella standard for product identification that encompasses UPC (12-digit, North America), EAN (13-digit, international), and ISBN (books). GTINs allow AI shopping agents to unambiguously match products across merchants, verify product authenticity, and aggregate reviews. AgentReadyHQ audits consistently find that 87% of non-Shopify stores lack GTINs in their structured data, making it the single most common issue blocking AI commerce readiness.
I
Instant Checkout
ChatGPT's one-click purchase feature powered by Stripe ACP, charging 4% commission.
Instant Checkout is the native transaction feature within ChatGPT Shopping that allows consumers to purchase products without leaving the chat interface. Powered by Stripe through the ACP protocol, it stores payment information securely and enables one-click buying. Merchants pay a 4% commission on each Instant Checkout transaction in addition to standard Stripe processing fees. Setting up Instant Checkout requires Stripe account integration and ACP protocol enablement.
ISBN (International Standard Book Number)
A GTIN variant used specifically for books and publications.
ISBN is a unique numeric identifier assigned to each edition of a book or book-like publication. Since 2007, ISBNs use a 13-digit format that is compatible with the EAN/GTIN system. For bookstores and publishers selling through AI shopping channels, including ISBNs in product structured data is essential for accurate product matching and cross-merchant comparisons by AI agents.
J
JSON-LD (JavaScript Object Notation for Linked Data)
The recommended format for embedding schema markup in web pages.
JSON-LD is a lightweight data serialization format that embeds structured data directly into the HTML of web pages using script tags. It is the format recommended by Google and preferred by AI agents for reading schema.org markup including Product, Offer, BuyAction, and Organization types. Unlike microdata or RDFa, JSON-LD is decoupled from the visible HTML, making it easier to implement and maintain without altering page templates.
L
LLMs.txt
A standardized file at your domain root that tells AI models about your store.
LLMs.txt is an emerging standard for a plain-text file placed at the root of your domain (e.g., yourstore.com/llms.txt) that provides AI language models with a structured overview of your business. It describes what you sell, your areas of specialization, how to navigate your catalog, and any relevant policies. Think of it as a robots.txt for conversational AI: it helps language models understand and accurately represent your store when generating shopping recommendations.
M
MCP (Model Context Protocol)
Anthropic's open protocol letting AI models connect to external tools and data.
MCP is an open protocol created by Anthropic that allows large language models to securely connect to external data sources and tools in real time. For e-commerce, MCP enables AI agents to query live inventory, check current pricing, and initiate transactions through standardized interfaces. MCP has been endorsed by Google, Microsoft, and OpenAI, positioning it as a foundational layer for agentic commerce alongside UCP and ACP.
Merchant Center
See: Google Merchant Center.
Merchant Center is the commonly used shorthand for Google Merchant Center, the platform where merchants manage their product data feeds for Google Shopping and Google AI Mode. See the full Google Merchant Center entry for details.
Multimodal Search
AI search that processes images, text, and voice simultaneously.
Multimodal search allows consumers to use combinations of images, text, and voice in a single query. For example, a shopper can photograph a piece of furniture and ask the AI to find a similar item in a different color. Amazon reports a 6% increase in click-through rate from multimodal AI search, while eBay reports a 31% increase in purchase rate. Merchants should ensure their product images are high-quality, properly tagged with alt text, and accompanied by comprehensive text descriptions to be surfaced by multimodal search.
O
Offer Schema
Schema.org markup describing a product's price, availability, and purchase conditions.
Offer schema (schema.org/Offer) is the structured data type that encodes the commercial details of a product listing: price, currency, availability status, condition (new/used), seller information, and valid date ranges. It is typically nested within Product schema. AI shopping agents rely heavily on Offer data to compare options across merchants, surface in-stock items, and present accurate pricing to consumers.
P
Perplexity Shopping
Perplexity AI's shopping feature with citation-based product recommendations.
Perplexity Shopping is the commerce feature within Perplexity AI that recommends products alongside detailed citations and source links. Unlike ChatGPT Shopping, Perplexity currently charges no commission fees and focuses on information-rich product discovery rather than in-chat transactions. Merchants benefit from Perplexity Shopping by having comprehensive, well-cited product content that the AI can reference when generating recommendations.
Product Feed
See: Data Feed.
Product feed is a synonym for data feed, referring to the structured file containing your product catalog information that is submitted to shopping platforms and AI commerce services. See the full Data Feed entry for details.
Product Schema
Schema.org markup describing a product's properties to search engines and AI agents.
Product schema (schema.org/Product) is structured data that describes all attributes of a product: name, description, images, brand, SKU, GTIN, price, availability, reviews, and ratings. It is the foundational schema type for e-commerce and is embedded in pages using JSON-LD. AI shopping agents parse Product schema to understand what you sell, how much it costs, and whether it is in stock. Incomplete Product schema is one of the most common reasons stores score low on AI readiness audits.
R
Rich Results
Enhanced search result displays powered by structured data.
Rich results are visually enhanced search listings that display additional information such as star ratings, pricing, availability badges, and product images. They are triggered by correctly implemented structured data on your pages. While rich results have traditionally been a Google Search feature, the same structured data that powers them is also used by AI agents to generate accurate product recommendations. Implementing rich-result-worthy schema is therefore a two-for-one investment in both traditional and AI search visibility.
Robots.txt
A file controlling web crawler access. Must not block AI agent crawlers for AI visibility.
Robots.txt is a plain-text file at your domain root that instructs web crawlers which pages they may or may not access. For AI commerce, it is critical that your robots.txt does not block the user agents of AI shopping crawlers, including GPTBot (OpenAI), Google-Extended (Gemini), PerplexityBot, and ClaudeBot (Anthropic). Blocking these crawlers makes your products invisible to the corresponding AI shopping platforms, regardless of how well your structured data is implemented.
S
Schema Markup
Structured data vocabulary helping search engines and AI agents understand page content.
Schema markup refers to the vocabulary defined at schema.org that provides a standardized way to annotate web content for machine consumption. For e-commerce, the most important schema types are Product, Offer, BuyAction, Organization, and BreadcrumbList. AI agents depend on schema markup to accurately interpret what your pages contain without needing to parse free-form HTML. Research from Princeton shows that GPT-4's product data accuracy jumps from 16% to 54% when structured data is present.
Structured Data
Machine-readable information about your products and store.
Structured data is any information formatted in a standardized, machine-readable way. In e-commerce, this includes schema markup embedded in HTML, product data feeds (XML/CSV/JSON), and API endpoints that return formatted product information. Structured data is the foundation of AI commerce visibility. Research from Princeton demonstrates that GPT-4's accuracy in extracting product information rises from 16% to 54% when structured data is present, making it one of the highest-impact optimizations a merchant can make.
T
Transactability
The ability for AI agents to complete purchases on your store.
Transactability measures whether an AI shopping agent can go beyond recommending your product and actually complete the transaction. Achieving transactability requires implementing a commerce protocol (UCP for Google, ACP for ChatGPT), connecting a supported payment processor (typically Stripe), and adding BuyAction schema markup so agents know the purchase endpoint. Without transactability, your store can be discovered by AI agents but not sold through them, capping your AI commerce potential at referral traffic only.
U
UCP (Universal Commerce Protocol)
Google's open protocol for enabling AI agents to discover and purchase products.
UCP is an open commerce protocol developed by Google that provides a standardized interface for AI agents to discover product catalogs, check inventory, and process transactions. It has been endorsed by BigCommerce, Adobe Commerce, and other major platforms. UCP currently carries 0% commission, making it the most cost-effective AI commerce channel. Merchants who implement UCP gain access to Google AI Mode's native shopping experience and any other platform that adopts the protocol.
UPC (Universal Product Code)
A 12-digit product barcode identifier used primarily in North America.
UPC is the 12-digit product identification standard used predominantly in the United States and Canada. It is the most common form of GTIN in the North American market and appears as the familiar barcode on physical product packaging. AI shopping platforms use UPCs to match products across different merchants, ensuring that comparison results are accurate. Including UPCs in your Product schema and data feeds is essential for AI commerce readiness.
W
Waiting Tax
The compounding trust deficit stores accumulate with AI agents by delaying optimization.
The Waiting Tax is an AgentReadyHQ concept describing how the cost of inaction compounds over time in AI commerce. AI agents build trust scores and preference models based on historical data quality, consistent availability, and protocol compliance. Merchants who optimize early accumulate trust that late movers cannot quickly replicate, similar to how domain authority in traditional SEO takes years to build. Every month of delay widens the gap between early adopters and those still waiting.
Z
Zero-Click Search
A search where users get answers directly without clicking through to a website.
Zero-click searches occur when a search engine or AI assistant provides a complete answer directly in the results interface, eliminating the need to visit an external website. SparkToro data shows that 60% of Google searches are now zero-click. AI shopping agents accelerate this trend by synthesizing product information from multiple sources into a single conversational response. Merchants must ensure their structured data is comprehensive enough to be represented accurately in these zero-click experiences, even when the consumer never visits their site.
Why This Glossary Exists
AI commerce moves fast. In the past six months alone, Google launched UCP, OpenAI launched ACP, and Anthropic opened MCP for commerce. Merchants are drowning in acronyms while trying to keep their stores competitive. This glossary is your reference point — bookmark it and come back whenever you hit a term you don't recognize. We update it as new protocols, platforms, and concepts emerge.
87%
of stores lack GTINs in structured data
11.4%
average AI-referred conversion rate
60%
of Google searches are zero-click
Ready to Put These Terms Into Action?
Get your free Agent Ready Score and see how well your store implements the concepts in this glossary.