Agentic Commerce for Online Merchants

AI agents are becoming the new sales channel. Merchants who are not visible to ChatGPT, Google Business Agent, and others will lose customers to competitors who are. Here is the roadmap.

Why Act Now?

ChatGPT has over 700 million weekly users. A growing share of these users researches and buys products through AI assistants. Google is building Business Agents directly into search. Shopify is preparing over one million merchants for ACP compatibility.

The question is not whether Agentic Commerce is coming, but when it becomes relevant for your industry. And the preparations — structured data, APIs — also improve your traditional SEO and your multichannel strategy.

Step 1: Structured Product Data

The most important and most urgent step. AI agents need machine-readable data:

• Schema.org JSON-LD on every product page: Product, Offer, AggregateRating, Review
• Complete attributes: Name, description, price, availability, variants, images, EAN/GTIN
• Clean categorization: Google Product Taxonomy or your own consistent hierarchy
• Real-time accuracy: Prices and availability must be up to date — incorrect data = agent stops recommending

Step 2: API-Capable Backend

Your shop must be able to accept orders via an API. Headless commerce architectures have a natural advantage here — the API already exists. For traditional shops, it means: Provide REST or GraphQL endpoints for product search, cart, and checkout.

Step 3: MCP Integration

The Model Context Protocol is the fastest way to become reachable for AI agents. An MCP server describes your shop's capabilities in a machine-readable way — agents automatically recognize what is possible (Tool Discovery). Shopify merchants already have this via the Shopify MCP server.

Step 4: Implement ACP or UCP

Which protocol you implement first depends on your ecosystem:

• Targeting ChatGPT? -> ACP. Use Stripe integration.
• Targeting Google? -> UCP. Use Google Merchant Center.
• Both? -> Start with the protocol that fits your existing infrastructure, then expand.

Step 5: Prepare Payment

Your payment provider must be able to process tokenized payments (SharedPaymentTokens for ACP, Mandates for AP2). Stripe, PayPal, and the major card networks are actively building these capabilities. More details: Payment in Agentic Commerce.

Step 6: Monitoring

Track agent activity: How many requests come from AI agents? Which products are queried most frequently? What is the conversion rate in the agent channel? This data helps optimize the product feed and API endpoints.

E-Commerce Platforms Compared

How well is your platform prepared for Agentic Commerce? The differences are significant — from native protocol support to custom development.

High Readiness

Medium Readiness

In Development

What Changes in Marketing

The paradigm shift in one sentence: From ads to reputation, from UX to Agent Experience.

• SEO becomes GEO: No longer just optimizing for Google rankings, but ensuring AI agents recommend your products
• Data quality beats design: An agent doesn't see a beautiful product page — it sees structured data. Those with better data get recommended more often.
• Reputation becomes measurable: Agents learn which shops are reliable. On-time delivery, accurate data, and hassle-free returns are directly rewarded.
• Advertising becomes less relevant: An agent decides based on data, not based on banners or sponsored placements.