Skip to main content

​
Letting AI Place Orders for Customers

The AI agent has found your products and trusts you. Next question: how does it actually help a consumer buy something from your store? Traditional checkout flows are designed for humans: click “Add to Cart,” fill in a shipping address, select a payment method, click “Confirm Order.” AI agents do not click buttons. They need a set of machine-to-machine communication protocols. In 2026, three protocols are defining how AI agents interact with merchants.

​
4.1 Three Protocols, Three Problems Solved

ProtocolProblem It SolvesCreated By
UCPHow does an AI agent browse your products?Co-developed by Google, Shopify, Etsy, Wayfair, Target, Walmart
ACPHow does an AI agent place an order for the consumer?Created by Stripe; OpenAI ChatGPT is the first implementer
MCPHow does AI become your sales assistant?Created by Anthropic, now managed by the Linux Foundation

​
4.2 The Storefront Analogy

Imagine you run a store, and out front you have installed three different devices:

​
UCP = The Display Window

A transparent display window. An AI agent walks by and can see what you sell, at what price, and whether it is in stock. It does not need to enter the store or speak with you. Through the glass, it takes your product information and presents it to the consumer. In technical terms: UCP defines a standard set of API endpoints. An AI agent calls these endpoints to retrieve your product listings, search products, and check inventory and prices.

​
ACP = The Self-Checkout Kiosk

The consumer says “I want to buy this.” The AI agent walks up to the self-checkout kiosk, scans the item, charges the consumer’s card, prints a receipt, and completes the transaction. No human intervention required. In technical terms: ACP defines the complete cart, checkout, payment, and order management flow. It integrates deeply with Stripe to ensure payment security.

​
MCP = The Smart Sales Associate

An AI-powered sales associate standing inside your store. Customers can ask anything, and the associate has the answers:
  • “Do you have this shoe in size 9?” — check inventory
  • “What is the difference between these two models?” — product comparison
  • “Where is the order I placed last week?” — track order status
In technical terms: MCP lets an AI assistant call “tools” and access “resources” that you provide. You define the capabilities; the AI assistant invokes them as needed.

​
4.3 How the Three Protocols Work Together

Consumer: "Find me a pair of waterproof running shoes"
    |
    v
AI agent uses UCP to browse product catalogs from multiple merchants
    |  -> Finds 3 stores with waterproof running shoes
    |
AI agent uses OTR to check trust scores for all 3 merchants
    |  -> Store A: GOLD (85), Store B: SILVER (72), Store C: BRONZE (61)
    |
AI agent recommends Store A and Store B to the consumer
    |
Consumer: "Buy the ones from Store A"
    |
    v
AI agent uses ACP to complete checkout at Store A
    |  -> Cart -> Address -> Payment -> Order confirmed
    |
Consumer: "Has it arrived yet?"
    |
    v
AI agent uses MCP to call Store A's order tracking tool
    -> "Your package has shipped. Estimated delivery: tomorrow"

​
4.4 Which Protocol Do You Need?

Not every merchant needs all three protocols. Choose based on your situation:

​
I just want AI to discover and recommend my products

-> Start with Schema.org + llms.txt + Sitemap (the essentials from Chapter 3) -> Consider adopting UCP so AI agents can browse your catalog directly

​
I want AI agents to place orders for consumers at my store

-> Adopt UCP (product browsing) + ACP (checkout and payment) -> Requires developer support, or use ORBEXA’s one-click integration

​
I want AI to act as my customer service / sales assistant

-> Adopt MCP and define your commerce tools -> Consumer AI assistants can then check inventory, track orders, and answer questions directly

​
I want it all

-> Full UCP + ACP + MCP integration -> ORBEXA provides a complete protocol translation layer — one integration covers all three
Practical advice: For most merchants, the highest-impact first step is the three essentials from Chapter 3 (Schema.org + llms.txt + Sitemap). Protocol adoption can be phased in gradually.

​
Deep Dives on Each Protocol

Each protocol has its own dedicated guide:

​
Self-Check Checklist

  • Understand what problem each of the three protocols solves
  • Identify which protocol you need most right now
  • If you are a Shopify merchant -> next chapter
  • If you are a WooCommerce merchant -> Chapter 6
  • If you run a custom-built site -> Chapter 7
Next chapter: Shopify Merchant Setup