Donate to support freedom.
Get the same
Vasilkoff: software development company in Cyprus

Build Commerce Agents with UCP: Sell Inside AI Channels

Turn Shopify into an Autonomous Agentic Storefront with Google and Shopify's Universal Commerce Protocol

Learn how to build commerce agents with Shopify and Google's Universal Commerce Protocol (UCP), and unlock high-intent sales in AI chat channels.


Time to read: 8 min

Editorial illustration showing an AI commerce agent communicating through the Universal Commerce Protocol to a verified merchant store.

From Browsing Catalogues to Autonomous Agentic Execution

Build Commerce Agents with UCP: How to Sell Inside AI Channels

Online retail is going through its most significant structural transformation since the advent of the smartphone.

For decades, e-commerce depended on a predictable consumer funnel: a shopper entered a search query into Google, clicked an ad or an organic link, browsed a catalogue, added items to a cart, and navigated a multi-step checkout form.

Today, that funnel is fragmenting. Consumers are increasingly turning to conversational AI assistants like Google Gemini, Microsoft Copilot, ChatGPT, and Perplexity to solve their shopping needs. Instead of browsing ten tabs, a buyer gives an agent a set of constraints:

"Find me a waterproof 40L hiking pack with a padded 16-inch laptop sleeve and ergonomic hip belt under $180 that delivers to my address by Thursday."

The AI doesn't just return links. It evaluates options, confirms inventory and return policies, builds a cart, and—with the shopper's approval—completes the transaction right inside the conversation.

This is agentic commerce. According to McKinsey, AI agents are projected to mediate $3 trillion to $5 trillion in consumer commerce by 2030. In the second quarter of 2026 alone, AI-driven traffic to Shopify merchants surged by 197% year over year, with orders originating from AI searches tripling.

To enable this transition without forcing merchants to build bespoke integrations for every emerging AI platform, Shopify and Google co-developed an open standard: the Universal Commerce Protocol (UCP).

In this guide, we break down what UCP is, how to build commerce agents that leverage it, and what this technological leap gives to Vasilkoff clients.

What is the Universal Commerce Protocol (UCP)?

The Universal Commerce Protocol (UCP) is an open-source standard designed to serve as the "universal translator" between autonomous AI agents and e-commerce backends.

Historically, connecting a store to external platforms required individual API integrations or fragmented plugins. If you wanted to sell in ChatGPT, Microsoft Copilot, Google AI Mode, and Meta AI, you faced a nightmare of custom APIs, divergent data structures, and brittle middleware.

UCP solves this by standardizing how AI agents interact with retail systems across four foundational pillars:

[ Shopper Intent ] 
        │
        ▼
[ AI Agent (Gemini / Copilot / ChatGPT) ]
        │  (Universal Commerce Protocol - UCP)
        ▼
[ Capabilities Negotiation ] ──▶ [ Product Discovery ]
        │
        ▼
[ Cart Creation & Validation ] ──▶ [ Secure Checkout & Settlement ]
  1. Capability Discovery: Standardizes how a merchant declares what its store can do (e.g., supported shipping zones, split deliveries, local pickup, currency conversions, and return rules).
  2. Catalog Discovery & Semantic Matching: Exposes structured, machine-readable inventory so AI models can verify real-time availability, specifications, and pricing without hallucinating.
  3. Cart & Transaction Negotiation: Allows the AI agent to assemble line items, apply valid discount codes, compute accurate taxes, and calculate shipping rates programmatically.
  4. Delegated Checkout & Settlement: Facilitates secure, one-click transactions using protocols like the Agent Payments Protocol (AP2) and Shop Pay, keeping sensitive credentials safe while leaving the merchant as the official Merchant of Record.

Backed by Shopify, Google, Microsoft, Meta, Stripe, and major retail enterprises, UCP ensures merchants don't surrender their customer relationships or brand sovereignty to walled gardens.

How UCP Works Under the Hood

To understand how to build commerce agents with UCP, consider how an agent executes a purchase on behalf of a shopper:

1. Intent Parsing and Catalog Query

When a customer prompts an AI assistant, the model identifies commercial intent and maps customer constraints against UCP-compliant endpoints. Instead of parsing messy HTML, the agent queries machine-readable endpoints:

{
  "protocol": "ucp/v1",
  "action": "query_catalog",
  "parameters": {
    "category": "backpacks",
    "attributes": {
      "capacity_liters": { "gte": 40 },
      "features": ["waterproof", "laptop_sleeve_16in"],
      "max_price": 180.00,
      "currency": "USD"
    },
    "fulfillment": {
      "destination_postal_code": "94103",
      "required_by": "2026-10-01"
    }
  }
}

2. Capabilities & Policy Validation

Before presenting an option, the agent verifies merchant policies. Does the store offer free returns? Are there restocking fees? Is delivery guaranteed within the requested window? UCP allows the store's knowledge base to answer these deterministically.

3. Cart Construction via Model Context Protocol (MCP)

The agent leverages an MCP server provided by Shopify or custom-built by developers. It issues a structured tool call to initialize a cart:

{
  "tool": "create_ucp_cart",
  "arguments": {
    "merchant_id": "merchant_78291",
    "items": [
      {
        "variant_id": "gid://shopify/ProductVariant/44910283",
        "quantity": 1
      }
    ],
    "shipping_address": {
      "country": "US",
      "zip": "94103"
    }
  }
}

4. Delegated Checkout (AP2 & Shop Pay)

Rather than passing raw credit card numbers through an LLM prompt (which creates severe PCI-DSS compliance risks), UCP uses delegated payment authorization.

The customer authenticates using biometric confirmation (such as Apple Pay, Google Pay, or Shop Pay) within the AI conversation. The merchant receives a cryptographically signed payment token and completes the transaction directly through their existing gateway.

How to Build Commerce Agents with UCP: The Technical Roadmap

Building commerce agents for your business isn't about training a massive language model from scratch. It is about engineering your store's architecture so AI agents can discover, verify, and transact against your inventory without friction.

Here is the step-by-step roadmap Vasilkoff implements for clients:

Step 1: Clean and Enrich Machine-Readable Product Data

AI agents do not browse websites like human eyes do. They do not get swayed by flashy hero sliders or vague marketing copy like "revolutionary comfort" or "next-generation style".

Agents search by hard technical specifications. If a customer asks for a "hypoallergenic bamboo bedsheet fitted for 14-inch deep mattresses," and those specs are trapped inside an unformatted image or an unstructured paragraph, the agent skips your product.

Best Practices:

  • Populate all standardized Shopify taxonomy attributes (dimensions, materials, weight, certifications, compatibility).
  • Implement robust JSON-LD Schema Markup (Product, Offer, AggregateRating, ShippingDetails, MerchantReturnPolicy).
  • Audit your robots.txt and ensure AI search crawlers (Google-Extended, GPTBot, PerplexityBot) are not accidentally blocked from indexing product metadata.

Step 2: Structure Your AI-Facing Knowledge Base

Autonomous agents must confirm merchant policies before closing a transaction. If a user says "Only buy from stores with hassle-free 30-day free returns", an agent will verify that rule before generating a checkout button.

  • Centralize return policies, warranty terms, and shipping rate tables in clear, crawlable plain language.
  • Avoid burying crucial terms inside JavaScript-heavy accordions or pop-up modals that agent parsers cannot evaluate.
  • Configure Shopify's native Knowledge Base and Agentic settings to ensure your official brand facts are indexed.

Step 3: Configure UCP and Model Context Protocol (MCP) Endpoints

Shopify has made stores agent-ready by default for eligible merchants through Agentic Storefronts (accessible under Sales channels > Agentic in the Shopify admin).

For brands requiring custom workflows—such as B2B trade discounts, bespoke product configurators, or cross-system ERP inventory verification—we build dedicated UCP-compliant MCP servers.

Using the Shopify AI Toolkit and UCP specifications, we declare tool capabilities that external agents can invoke:

  • search_products(query, filters)
  • validate_inventory(variant_ids, location)
  • calculate_custom_pricing(customer_segment, line_items)
  • generate_secure_checkout(cart_id, customer_token)

Step 4: Implement Secure Payment Delegation

Security is paramount. In an agentic environment, you must never allow an AI agent to handle raw customer financial credentials.

By integrating UCP with Shop Pay and the Agent Payments Protocol (AP2):

  • Customer payment methods are stored securely in certified vaults.
  • The agent passes a scoped intent request.
  • The shopper approves the final total with biometrics.
  • The order flows seamlessly into your Shopify Orders dashboard with full attribution.

Step 5: Test with Synthetic Shopper Agents & Monitor Attribution

Before going live, we run automated synthetic tests against conversational models:

  • Querying ChatGPT, Gemini, and Copilot with long-tail buyer prompts.
  • Verifying whether products surface in the top recommendations.
  • Ensuring checkout links properly authenticate and attribute sales back to the respective AI channel.

What UCP and Commerce Agents Give to Our Clients

Adopting UCP and agentic commerce is not just an experimental tech flex; it delivers tangible commercial advantages:

1. Direct Access to High-Intent, High-Converting Shoppers

Shoppers using AI agents aren't casually window shopping. They have specific problems and are ready to buy. Shopify's data shows that buyers arriving through AI channels convert at nearly twice the rate of traditional web channels.

By enabling UCP, your catalogue becomes immediately purchasable right when customer intent peaks.

2. A Massive Advantage for Niche and Specialized Products

In traditional Google Search or social media advertising, large enterprises with multi-million-dollar budgets dominate broad keywords.

Agentic commerce levels the playing field: 75% of AI-attributed sales on Shopify happen in categories outside the top 100. AI agents match by precision, not ad spend. If your product matches the customer's exact technical criteria, the agent selects your brand, regardless of whether you're a boutique manufacturer or a Fortune 500 brand.

3. Zero Operational Overhead Across Multiple AI Channels

Previously, expanding to a new sales channel meant managing separate inventories, disjointed pricing lists, and clunky third-party plugins.

With UCP, you list your products once in your central Shopify catalogue. That single source of truth syndicates simultaneously across:

  • Google AI Mode & Gemini
  • Microsoft Copilot
  • ChatGPT
  • Meta AI
  • Shop App

Every order lands in your unified Shopify dashboard with centralized inventory syncing and financial reporting.

4. Preservation of Brand Sovereignty and Customer Data

When merchants sell on third-party marketplaces like Amazon, they are often reduced to anonymous fulfillment providers. Amazon owns the customer email, the purchase data, and the post-sale relationship.

With UCP on Shopify, you remain the Merchant of Record. The customer buys from your brand, customer data flows into your CRM, and you retain full post-purchase marketing rights.

5. Future-Proofing for Agent-to-Agent (A2A) Commerce

We are rapidly approaching an economy where consumer agents talk directly to merchant agents. Soon, personal agents will autonomously restock household goods, negotiate bulk discounts for small businesses, and optimize shipping schedules without manual intervention.

Brands with UCP-compliant architecture are already positioned for this shift; those relying on legacy infrastructure will find themselves shut out.

The Foundation Matters: Why Modern SaaS Wins Over Legacy CMS

Building an agentic store on legacy, self-hosted platforms like WooCommerce, OpenCart, or older Magento builds is an uphill battle.

As we detailed in our analysis of Why We Choose Shopify for E-Commerce Growth, self-hosted systems suffer from "plugin hell", unstable API versions, and hosting vulnerabilities. When autonomous agents query your inventory thousands of times per minute during peak shopping surges, a fragile self-hosted server crashes.

Shopify’s global cloud infrastructure, 99.99% uptime, native PCI-DSS Level 1 compliance, and first-party UCP integration provide the rock-solid stability required for agentic commerce.

Partner with Vasilkoff to Build Your Agentic Commerce Engine

Transitioning your store into an agent-ready powerhouse requires more than ticking a checkbox. It demands:

  • Precision product data engineering and schema modeling
  • Custom theme optimization for lightning-fast crawler indexing (see our Custom Shopify Theme Development solutions)
  • Tailored UCP/MCP integrations for specialized business logic
  • End-to-end testing with senior software engineers

At Vasilkoff Ltd, we have been building accountable software solutions since 2012. We combine the rapid execution of modern AI tools with senior engineering oversight—and we back all our work with our free lifetime bug-fixing guarantee.

Ready to capture high-intent buyers in the age of AI agents?

👉 Contact our engineering team or talk with our AI project intake assistant today to map out your agentic commerce strategy.

Last updated: