Agentic Commerce FAQ & Glossary: Every Term Explained
Agentic commerce arrived with its own vocabulary, and a lot of it gets used loosely — “agent,” “entity,” and “zero-click” mean specific, different things depending on who’s talking. This glossary defines every core term you’ll run into across our Agentic Commerce & Agent Experience (AX) track in plain, practitioner language, grouped by theme so related concepts sit together. Bookmark this page; we link back to it from every other article in the series instead of re-explaining terms each time.
Where a term has a nuance that trips people up — like the difference between a “feed” and “schema,” or “AI Overviews” and “zero-click” — we call it out directly rather than leaving you to infer it.
Core Concepts
Agentic commerce
The buying and selling of products through AI agents acting on a person’s behalf — researching, comparing, and sometimes completing checkout — instead of a human clicking through search results and a storefront manually.
AI agent
Software, usually built on a large language model, that can interpret a goal, take multi-step action toward it (searching, comparing, calling other tools or services), and in some cases complete a transaction, with varying degrees of human oversight. A general-purpose assistant like ChatGPT or Claude becomes an “agent” in this sense the moment it starts taking actions rather than just answering a question.
Agentic checkout
A purchase flow designed to be completed by an AI agent rather than a human filling out form fields — typically using a structured protocol that passes product, price, and payment authorization data programmatically instead of relying on a human clicking buttons on a checkout page.
Agent Experience (AX)
The practice of making a website, product feed, and content legible and trustworthy to an AI agent, the same way UX makes a site usable for a person and SEO makes it findable by a search engine. AX asks: can an agent identify what you sell, verify your claims against structured data, confirm you’re credible, and complete or hand off a transaction without hitting a dead end.
Delegated purchase
The far end of the agentic commerce spectrum, where an agent completes a transaction autonomously within limits a human set in advance — a budget, a reorder trigger, a preferred brand list — without approving each individual purchase.
Data and Feeds
Product feed
A structured file — commonly XML or JSON — listing a merchant’s products with standardized fields like title, price, availability, GTIN/SKU, and images. Feeds were built originally for shopping ads, but agentic platforms increasingly pull from the same feeds (or a similar feed format) to know what you sell and whether it’s in stock right now.
Structured data / schema markup
Code embedded in a webpage, usually in JSON-LD format, that labels content in a machine-readable way — this number is a price, this text is a review rating, this name is the author. Structured data lets an agent extract facts with certainty instead of guessing from unstructured page copy. Schema.org is the shared vocabulary most structured data follows.
Entity
A distinct, identifiable “thing” — a person, business, product, or concept — that search engines and AI systems can recognize and connect facts to, independent of any one page’s wording. Building entity clarity (a consistent name, consistent attributes, and links between your site and authoritative references like your Google Business Profile or Wikipedia/Wikidata where relevant) helps an agent know that “Salterra Digital Services” mentioned in one place and “Salterra” mentioned in another are the same thing.
Canonical source of truth
The single, authoritative location for a fact about your business or product — your own site, not a third-party aggregator or an outdated marketplace listing. Agents that find conflicting information across sources tend to trust the source that’s most current, most structured, and most consistent with other verified data about the same entity.
Product attribute
A specific, comparable fact about a product — material, dimensions, weight, color, compatibility — expressed as discrete, structured data rather than buried in a paragraph of marketing copy. Attributes are what an agent actually compares across competing products; prose describing the same facts is far harder for it to extract reliably.
Protocols and Technical Infrastructure
MCP (Model Context Protocol)
An open standard that lets an AI model connect to external tools, data sources, and services in a consistent way, instead of every integration being custom-built. In plain terms: MCP is a common plug shape so an AI assistant can “plug into” your product catalog, booking system, or database without a bespoke integration for every AI platform it talks to. You don’t need to build MCP support yourself to benefit from agentic commerce, but understanding it explains why standardized, structured data matters more every year — protocols like this are what make your data usable by an agent in the first place.
Agent protocol
A broader term for any standardized way an AI agent communicates with an external system — placing an order, checking inventory, authorizing a payment. Agentic checkout protocols are one category; MCP is another, aimed more generally at tool and data access.
API (Application Programming Interface)
A defined way for one piece of software to request data or trigger an action in another, without a human interface in between. Agents rely heavily on APIs — a retailer’s inventory API, a payment API — to act rather than merely describe.
Function calling / tool use
The mechanism by which an AI model decides mid-conversation to invoke an external tool — a search, a calculator, a checkout API — rather than answering from its own training alone. This is the technical behavior underneath most “agentic” capability: the model recognizes it needs current, real-world data or needs to take an action, and calls the appropriate tool to do it.
Retrieval
The step where an AI system pulls candidate information — web pages, product listings, database records — before generating a response. Retrieval quality depends heavily on how crawlable, structured, and current your content is; an agent can’t retrieve what it can’t parse or find.
Visibility and Measurement
Share of model
How often your brand, product, or content gets surfaced, cited, or recommended by AI systems relative to competitors, across the range of prompts a real customer might use. It’s the AI-era analog of share of voice or share of search — except instead of measuring rankings on a results page, you’re measuring how consistently you show up inside AI-generated answers and recommendations.
AI Overviews
The AI-generated summary Google shows above traditional search results for many queries, synthesizing information from multiple sources into a single answer with citations. Appearing as a cited source in an AI Overview is a distinct visibility goal from ranking #1 organically — the content that earns citation tends to be clear, well-structured, and directly answers the question asked, rather than being merely comprehensive.
Zero-click search
A search where the user gets their answer directly on the results page (or inside an AI-generated summary) and never clicks through to any website. Zero-click behavior has grown steadily and is a defining feature of the AI Overview and AI-assistant era — it means visibility and citation increasingly matter as much as, or more than, the click itself.
Citation / mention
An instance where an AI system references your brand, product, or content as a source when answering a query — with or without a clickable link. Unlike a traditional backlink, a citation inside a generated answer can influence a purchase decision even if no click ever happens, which is part of why “ranking” alone no longer captures the full picture of visibility.
Answer engine
An informal term for AI systems — chat assistants, AI Overviews, AI-powered search — that respond to a query with a synthesized answer rather than a list of links. “Answer Engine Optimization” (AEO) is sometimes used alongside AX and SEO to describe optimizing specifically for how these systems select and phrase what they surface.
Trust and Authority Signals
E-E-A-T
Experience, Expertise, Authoritativeness, and Trustworthiness — Google’s framework for evaluating content quality, now equally relevant to how AI systems weigh which sources to trust and cite. Content written by a named, credentialed person with demonstrable real-world experience tends to be favored over anonymous or thin content, in both traditional and AI-driven search.
Author markup
Structured data identifying who wrote a piece of content, often linked to a broader author profile establishing their credentials and other work. This gives both search engines and AI agents a verifiable signal of expertise behind a claim, rather than an unattributed assertion.
Review schema
Structured data marking up customer reviews and ratings in a machine-readable format — reviewer, rating value, date, review text — so an agent can factor genuine review sentiment into a recommendation instead of relying on a page’s marketing claims alone.
Frequently Asked Questions
What's the difference between a product feed and structured data?
A product feed is a separate file (usually XML or JSON) submitted directly to a platform listing your catalog, while structured data is code embedded in your webpage itself; both serve the same underlying goal of giving machines reliable, structured facts, and platforms increasingly draw on both.
Is MCP something I need to build myself?
No — MCP is infrastructure that AI platforms and developers implement; as a business owner, your job is making sure your own data (product info, inventory, content) is structured and accessible enough that any protocol built on top of it, MCP included, can use it accurately.
Does "share of model" replace keyword rankings as a metric?
It doesn't replace rankings, it adds to them — keyword rankings still matter for traditional search traffic, but share of model captures a separate and growing channel where visibility happens inside AI-generated answers rather than on a results page.
If AI Overviews cause zero-click searches, why bother ranking or getting cited at all?
Because citation still drives brand recall, trust, and eventual purchases even without an immediate click — being named as the recommended or cited source in an AI answer shapes the decision the same way a trusted referral does, whether or not that specific interaction produces a click.
Do I need agentic checkout set up before I can benefit from AX?
No — most of AX (clean structured data, accurate feeds, clear entity information, genuine reviews) improves your visibility to agents in the research and comparison phase regardless of whether checkout itself is automated, and that phase is where most agentic commerce activity currently happens.
Where should I start if this is all new terminology?
Start with structured data and your product feed — those two fixes give any AI agent the clearest, most reliable read on what you sell, and they're the foundation every other tactic in this track builds on.
Terry has 30+ years in software and SEO. He’s the founder of Salterra Digital Services and SEO Spring Training, host of the Roundtable SEO Mastermind, and lead instructor at SEO University — teaching the exact tactics his team uses on client work.
Continue learning
Ready to master this?
This guide is one lesson from the Agentic Commerce & Agent Experience Optimization (AX) course. Get every lesson, framework and checklist — plus the full 38-course catalog — inside SEO University.
Practitioner-focused training across the full digital marketing stack — from technical SEO to conversion optimization and the AI search era. By Salterra Digital Services, since 2011.