Google lança UCP, padrão aberto para viabilizar “agentic commerce” e compras dentro de experiências de IA

Google announced the Universal Commerce Protocol (UCP), a open standard and open-source designed to enable the next generation of agentic ''the scenario where AI agents not only recommend products, but also succeed perform steps of the purchase process from end to end. The UCP establishes a common“ and functional primitives to connect” consumption surfaces (as conversational experiences), or, business (retailers/marketplaces) and payment providers, while maintaining compatibility with existing retail infrastructure. Source

The protocol was developed in collaboration with leaders of the 'OD ecosystem with mentions of Shopify, Etsy, Wayfair, Target and Walmart, in addition to a wide range of partners such as Adyen, American Express, Mastercard, Stripe, Visa, Zalando and others. Source Source

What is UCP (in simple terms)

UCP standardizes capabilities (capabilities) essential to digital commerce (AS checkout, discounts and fulfillment ^^___________________________________________________________________________________________________________________________ discover and trigger these capabilities consistently, without “ integrations measured with” for each new interface/agent. Source

According to Google, the goal is to reduce the classic bottleneck of integrations “N×N” (each retailer integrating with each surface/agent separately) for a model with a standardized point of integration. Source

First applications: checkout in AI Mode and Gemini

Google claims that UCP will support a new checkout flow in eligible listings in the AI Mode (Search) and gemini App, allowing the user to complete purchases during the search (reducing friction and abandonment). Payment must occur via Google Pay, using information already saved in the Google Wallet, with mention of future support to PayPal. Source

Interoperability: why it matters technically (and strategically)

UCP is designed to be interoperable with standards and protocols linked to the agent layer including AP2 (Agent Payments Protocol), A2A (Agent2Agent) and MCP (Model Context Protocol) nd support multiple“ ”transports (APIs, MCP and A2A). In practice, this reduces lock-in and accelerates adoption by different platforms and stacks. Source Source

How retailers can start (view Google)

Google documentation indicates the Merchant Center as a way to participate in the implementation in the Google ecosystem, with integration options such as:

• Native checkout (standard)
• Embedded checkout (optional and for approved cases requiring high customization) Source

Market Analysis (Marketing & Business) 0 Wilson Silva

For Wilson Silva's — Professor of ESPM & SP, Coordinator of the MBA in Marketing and AI of Impacta, speaker in the area of AI and CEO WS Labs ''UCP represents a structural change in performance marketing and and e-commerce:

“O UCP repositions e-commerce: exits the model (click & click & & & (checkout‘ and enters the model ¡intention ̄s talk ̄s action ̄. Whoever masters this transition will first capture more conversions with less friction ¡and will reduce acquisition cost by shortening the path between discovery and purchase.’
— Wilson Silva's

In Wilson's view, new competitive advantage it ceases to be only media and creative and becomes “purchase readiness by” agents, on three fronts:

1. Structured data and offer for AI (being “understandable by” agents)
   Retailers will need to enrich product attributes and context (including answers to common questions, compatibilities, substitutes, etc.), as discovery on conversational surfaces ceases to be just keyword and becomes semantics + intention. Source
2. Conversion as a product (not as a final step)
   Checkout within context reduces friction. This changes the role of e-commerce: the product page loses exclusivity as the“ decision ”place, and the decision can occur within the AI interface, provided the retailer is technically ready. Source
3. Post-purchase as differential (retention and LTV)
   If AI makes it easier to buy“, the war migrates to experience: delivery, exchanges, support, and loyalty. Retailers who treat post-purchase as a growth engine (rather than as cost) tend to earn LTV and reduce effective CAC.

How Brazilian companies should prepare in 90 days (full checklist)

015 Days ¡Diagnosis and database

• Feed/catalog audit: identify attribute gaps (descriptions, variations, policies, inventory).
• Standardize product taxonomy (categories, critical attributes, compatibilities).
• Map the 30 most frequently asked questions customer by category (to turn around “Decision content”).
• Define governance: who approves price, policy, promo, substitutions and discount rules.

1545 days (Preparar ao operation para “purchases by AI”

• Review checkout and policies (time, freight, return, chargeback) to reduce friction that agents cannot “negotiate”.
• Setting trade rules (e.g., when to allow replacement, when to offer bundle, when to give discount).
• Connect inventory in real time (where possible) to avoid breakup and frustration.

4590 Days & Pilots, metrics and scale

• Rotate pilot by category (e.g., 12 categories with better margin and simple logistics).
• Create “Agentic Readiness” metric” (e.g., data completeness, inventory SLA, return fee, delivery time).
• Review media strategy: from “clique” to “intention” campaigns and creatives go on to fuel conversation and decision, not just traffic.
• Training attendance for post-purchase (exchange, questions, delays) with a focus on reputation and retention.