Agentic Commerce Optimization (ACO) & UCP Protocol

1. Definition and Context

Agentic Commerce Optimization (ACO) is the strategic practice of structuring product data, signals, and technical infrastructure to ensure selection by autonomous AI agents. As the web transitions from human-led search to AI-mediated selection, ACO focuses on data integrity and machine-readability rather than visual persuasion.

This shift is driven by the Universal Commerce Protocol (UCP), a framework that expands commerce capabilities beyond simple checkout into discovery, loyalty, and post-purchase support. Unlike the Agentic Commerce Protocol (ACP), which focuses primarily on the transactional layer (Checkout → Fulfillment → Payment), UCP covers the entire commerce lifecycle, minimizing the “fragmented commerce journey” by allowing agents to integrate once and interact with multiple platforms seamlessly.

Current Deployment Status

UCP is currently active. As of February 2026, major platforms including Wayfair and Etsy have integrated UCP to enable direct purchasing within Google’s AI Mode.

2. The Six Core Capabilities of UCP

UCP defines six layered capabilities that agents use to navigate the commerce lifecycle. Optimization efforts must address each layer to ensure full agent compatibility.

1. Product Discovery: How agents find, parse, and surface inventory during the research phase.
2. Cart Management: Handling multi-item baskets, dynamic pricing, and complex basket rules.
3. Identity Linking: OAuth 2.0 authorization for personalized experiences and loyalty program integration.
4. Checkout: Session creation, tax calculation, and payment processing.
5. Order Management: Webhook-based lifecycle updates and logistical tracking.
6. Vertical Capabilities: Extensible modules for specialized use cases (e.g., travel booking windows, subscription schedules).

3. Technical Requirements: Schema & Data Structure

While UCP utilizes its own versioned JSON schema for transactions, Schema.org remains the critical “glue” for discovery. Agents use standard schema to decide who to transact with before UCP handles how the transaction occurs.

3.1 Mandatory Product Schema Fields

To ensure agent readability, the following fields must be populated with complete accuracy:

• Core Identity: name, description, SKU, GTIN, brand, image.
• Offers: Must include price, priceCurrency, availability, url, and seller.
• Trust Signals: aggregateRating and review are essential for third-party validation.
• Logistics: shippingDetails with precise delivery estimates.
• Variants: All product variants (size, color, material) must be explicitly defined.

3.2 Organization and Support Schema

• Merchant of Record: Use Organization (fallback to Person) to verify the entity responsible for the transaction.
• FAQPage: A designated FAQPage schema is required to support agent decision-making logic regarding policies and product details.

4. Merchant Center Configuration

UCP utilizes the Google Merchant Center feed as its primary discovery layer. Optimization requires specific attributes beyond standard feed requirements.

4.1 Critical Feed Attributes

• native_commerce: This attribute must be asserted to declare a product eligible for agentic checkout.
• Product Identifiers: Strict correlation between the feed ID and the product ID used in the Checkout API is required.
• Consumer Notices: Any product warnings must use the consumer_notice attribute.

4.2 Policy and Support Data

• Return Policies: Complete data on return costs, windows, and policy links is required for the agent to validate the merchant as a “Merchant of Record.”
• Customer Support: Structured support data allows agents to handle entry-level queries (L1 Support) autonomously, reducing human support load.

Implementation Note: Google recommends adding these UCP-specific attributes via a Supplemental Feed to prevent formatting errors in the primary feed.

5. Conversational Commerce Attributes

To minimize hallucination during the discovery phase, UCP introduces “Conversational Commerce Attributes.” These data points allow agents to parse specific product relationships and details that standard keywords miss.

5.1 Key Attribute Categories

• Compatibility: Explicitly defining what other products or systems the item works with (e.g., “Compatible with iPhone 15 Pro”).
• Substitution: Defining acceptable alternatives for out-of-stock scenarios to prevent cart abandonment.
• Related Products: Structured data for cross-selling opportunities.

5.2 Granular Detail (The “Wolf” Principle)

Agents require high-resolution descriptors. Generic attributes (e.g., Color: “Purple”) should be supplemented with specific, agent-parseable details (e.g., Color: “Wolf” or “Dark Slate”) to satisfy specific long-tail queries.

6. Strategic Implications: Multi-Modal Fan-Out

Optimizing for ACO increases visibility in Fan-Out Queries—where a single user prompt (e.g., “Plan a camping trip”) decomposes into multiple sub-intents (Tent, Sleeping Bag, Cooking Gear).

• Visual Fan-Out: Agents may analyze a single image and “fan out” to find all constituent products.
• Attribute Priority: Agents prioritize products with “Conversational Attributes” (Compatibility, Substitution) when assembling complex bundles from a single prompt.

7. Future Roadmap & Vertical Expansion

The UCP roadmap indicates expansion beyond retail into service-based verticals.

• Complex Baskets: Native support for bundling, promotions, and multi-item carts.
• Loyalty Integration: Standardized linking for points and member pricing.
• Post-Purchase Agents: Agents capable of managing returns and tracking without human intervention.
• New Verticals: Travel, Digital Goods, and Food/Restaurant services.