Explainer

How agentic commerce actually works.

A shopper asks an assistant what to buy. Somewhere between that question and a delivered order sit four agent channels, a protocol stack, and seven readiness dimensions. This page walks the whole path.

Run the Free Audit How the score works

The landscape

One brand, four buying surfaces.

Each surface runs different protocols and signals. Coverage is not one integration — it is a decision about which surfaces come first.

Live UCP
Google AI Mode
Agents browse, compare, and complete checkout inside the AI interface.

Your brand needs

UCP manifest, product feeds, Google Merchant Center, structured data

Live MCP / Plugins
ChatGPT Shopping
ChatGPT recommends and links to products; MCP tool-use enables direct transactions.

Your brand needs

GEO-optimized content, structured data, product feeds, MCP tools

Live Product Feeds
Perplexity Shopping
Recommendations arrive with citations. Rich structured data decides who gets cited.

Your brand needs

Rich product data, reviews, competitive pricing, trust signals

Emerging A2A / Custom
Enterprise AI Agents
Procurement agents source, compare, and buy for organizations via A2A and custom APIs.

Your brand needs

A2A agent cards, API exposure, machine-readable policies, B2B pricing

Check Readiness Across All Four

How it works

Chat request to delivery.

A conversation becomes a completed purchase without the customer visiting your site.

Personal AI Shopper

Watch how UCP automates the journey from a simple chat request to final delivery.

User Intent
Find me a great camera under $500!
Searching across merchants for the best cameras...
DEEP DIVE

How the Protocol Stack Connects Everything

Click any node to explore how the agentic commerce layers connect. Each component plays a critical role in enabling AI-powered commerce.

Consumer
Surfaces
UNIVERSAL COMMERCE PROTOCOL

Capabilities & Extensions

Product Discovery
Search & Filter
Cart
State Management
Identity Linking
Account Mapping
Checkout
AP2 Extension
Order
Lifecycle
Vertical Caps
Custom Modules

Underlying Communication & Transports

Standard APIs
REST / OpenAPI
Model Context Protocol
LLM Native
Agent2Agent Protocol
Autonomous Handoff
Business
Backends

Our framework

Seven dimensions of agentic readiness.

Most brands treat this as one protocol project. A gap in any one dimension caps what the other six can return.

01
AI Discoverability
Structured data, schema markup, Knowledge Graph presence, content built for generative engines.
02
Protocol Readiness
UCP for Google AI Mode, ACP and MCP for tool-using agents, A2A for agent-to-agent commerce.
03
Data Quality
Feed completeness, attribute accuracy, inventory sync, pricing consistency, variant mapping.
04
Transaction Enablement
Checkout APIs, payment handler exposure, cart management, secure session handling.
05
Trust & Verification
Review aggregation, merchant verification, return policy clarity, compliance signals.
06
Post-Purchase Experience
Order status APIs, tracking integration, return handling, customer communication.
07
Measurement & Optimization
Agent session attribution, conversion analytics, channel performance.
— Score all seven, free →
BEYOND STANDARD RETAIL

New Commerce Paradigms

Agentic commerce enables scenarios that were previously impossible due to fragmentation.

Retail Media

Contextual Commerce

A user watching a cooking video can buy the exact blender used by the creator directly within the video player overlay.

UCP Embedded Interface
Virtual Goods

Phygital Gaming

Gamers purchase a physical limited-edition hoodie for themselves and unlock a digital twin for their avatar in one transaction.

UCP Identity Linking
Smart Home

Autonomous IoT

A smart fridge detects low milk, queries 3 local grocers for the best price/delivery time, and places the order automatically.

UCP Agent Negotiation
Travel & Dining

Multi-Vertical Trip

An agent books a flight, reserves a hotel room, and secures a dinner reservation in a single atomic transaction.

UCP Cart Synchronization