Universal Commerce Protocol (UCP) and Agentic Commerce

Google has introduced the Universal Commerce Protocol (UCP), a new open standard designed to ensure that products can be better understood and recommended in Search, Shopping results, and AI Mode in the future. For us, this means taking a closer look at UCP and comparing it with OpenAI’s approach, the Agentic Commerce Protocol (ACP).

What is the Universal Commerce Protocol?

Together with several partners (including Shopify and Stripe), Google has introduced an open-source protocol intended to standardize how product information is captured, structured, and understood across the entire shopping ecosystem. Specifically, product data will be processed through a single, unified protocol, and this shared product understanding will power Search, Shopping, and AI-driven experiences.

This has several implications for product feeds and e-commerce in general, as it changes how products qualify for visibility in Search, Shopping, or AI Mode. In other words, the quality of your feed will determine whether your products are considered in AI-generated responses.

The attributes within the feed are especially important: the more detailed and consistent they are, the better AI Mode can compare products and provide recommendations. Structured data also remains a crucial factor, as it continues to support product discovery.

What does this mean in practice?

Customers with high purchase intent can convert directly within Google Search, AI Mode, or Shopping.

Example:

Jan is looking for a suitcase. For his research, he uses Google’s AI Mode and asks: “I need a compact, small suitcase that can be used as carry-on luggage on airplanes.” AI Mode searches, reviews product feeds from various retailers, and then recommends a selection of products. Jan chooses one, pays via Google Pay, for example, and completes the purchase directly within AI Mode using UCP. The retailer remains the contractual partner—not Google. This means the retailer retains control over Jan’s data but uses Google’s service to process the transaction.

Benefits of UCP

According to Google, the protocol offers the following advantages:

• For businesses:
  Products can be displayed at multiple touchpoints along the customer journey, such as during research in AI Mode or within the Gemini app. The company remains the “merchant of record” and retains full flexibility to tailor the checkout experience to its own requirements.

• For AI platforms:
  With standardized APIs, implementing Agentic Commerce for a specific target audience becomes straightforward, regardless of the AI environment.

• For developers:
  UCP is and will remain open-source, designed to be driven forward by an active community.

• For payment platforms:
  Due to UCP’s modular structure, offering a wide range of payment methods becomes simple. Security is a core focus, and every purchase includes cryptographically encrypted user consent.

• For buyers:
  UCP aims to eliminate friction throughout the customer journey, enabling seamless purchases—even when using membership benefits—at their favorite stores.

Special Features of UCP

Google highlights the following characteristics:

UCP is designed to be modular and extensible, supporting comprehensive and rapid commerce implementation through functions and extensions. Features and extensions can be selected to best meet development requirements (e.g., APIs, MCP, or A2A). Through tokenization, UCP enables retailers to offer secure and seamless payments using existing integrations with payment providers.

UCP is based on open standards and is fully compatible with relevant existing protocols such as A2A, AP2, or MCP.

What about user data?

The retailer remains the contractual partner for all transactions and retains full ownership of customer relationships and data. Google’s standard security and privacy practices apply, but the retailer remains responsible for the transaction.

Comparison with OpenAI’s Agentic Commerce Protocol (ACP)

OpenAI’s Agentic Commerce Protocol (ACP) follows a different approach: it focuses on AI agents that make purchases on behalf of users. In this case, the user instructs an AI agent to buy a suitable carry-on suitcase, adds further criteria, and the AI agent makes a decision based on those criteria. After the user provides final confirmation, the purchase is completed.

This also means that as a retailer, you do not necessarily have to choose between ACP and UCP. Ideally, you would support both, as the approaches differ slightly.

What is Agentic Commerce and what are AI agents designed for?

Simply put, AI agents allow tasks—and especially decisions—to be delegated to an AI-driven program. Based on the user’s specifications and constraints, this program completes tasks such as making online purchases.

Example:

Jan wants to buy new running shoes. He tells his AI agent:
“I need new running shoes. They should be comfortable and lightweight. I wear size 44. I don’t want to spend more than €100. I don’t have a brand preference.”

The AI agent searches online, compares shoes from different providers, makes a decision for Jan, and completes the purchase. Jan only needs to confirm the payment with the payment provider.

Agentic Commerce essentially describes this process: the AI agent acts as the buyer and executes the user’s request. The agent interacts with online shops, price comparison platforms, and review portals to make an informed decision. Early implementations already exist in the U.S., such as Amazon’s “Buy for Me” or Perplexity’s “Buy with Pro.”

ACP vs. UCP Comparison

Open Questions and Risks of Agentic Commerce

Several open questions remain and will need to be clarified during beta testing in the U.S.:

• Who is responsible for product selection? A ranking system will likely be required, but the ranking factors are not yet clear. Sponsored products and advertising will also play a role. OpenAI has announced that ChatGPT will soon include ads.

• Currently, agents can only purchase one product at a time. The same appears to apply within UCP. It remains to be seen whether this limitation will persist.

Agentic Commerce is not without risks. For example, systems can be vulnerable to prompt injection attacks. Additionally, users must manually confirm credit card requests. Time will tell whether this poses a security risk. More broadly, the question arises: how do AI agents make the right decision and select trustworthy, secure shops rather than simply choosing the lowest price?

Conclusion

At present, both OpenAI and Google offer waiting lists for participation. In Europe, we will likely have to wait a bit longer before these solutions become widely available. Until then, the key is to monitor developments and respond early.

Our experts at arboro are ready to support you on this journey. Get in touch with us—together, we will prepare your Shopware store for Agentic Commerce and the next stage of evolution in e-commerce.

Sources:

• https://www.youtube.com/watch?v=pCkn2gW7Zhg - Google stellt UCP vor: Ist das DER Game Changer? [Search Camp 412] von Bloofusion Germany

• https://www.heise.de/news/Universal-Commerce-Protocol-Neuer-Standard-fuer-Online-Shopping-11137804.html

• https://developers.googleblog.com/under-the-hood-universal-commerce-protocol-ucp/

• https://developers.google.com/merchant/ucp?hl=de

• https://www.reddit.com/r/AI_Agents/comments/1qap7ym/google_just_dropped_ucp_the_biggest_shift_in/

• https://www.checkout.com/blog/openai-acp-google-ucp-difference

• https://www.searchenginejournal.com/agentic-commerce-what-seos-need-to-consider-acp-ucp/563503/

• https://www.synthcommerce.com/blog/ucp-vs-acp-comparison

• https://www.nudgenow.com/blogs-copy/openai-acp-google-ucp-difference

• https://www.linkedin.com/posts/arthur-bedel_openai-%F0%9D%90%80%F0%9D%90%82%F0%9D%90%8F-%F0%9D%90%AF%F0%9D%90%AC-google-%F0%9D%90%94%F0%9D%90%82%F0%9D%90%8F-two-activity-7416480750453039104-zqEs/

• https://www.checkout.com/de-de/blog/agentic-commerce

• https://www.productsup.com/de/blog/was-googles-ucp-fuer-ecommerce-ki-shopping-und-produktfeeds-bedeutet/