With its latest announcement, Google is introducing a set of new tools designed to help advertisers ensure their products remain visible and competitive as traditional search continues to decline. While Google Ads campaigns were already able to surface within conversations in Gemini and AI Mode, this new suite of tools significantly expands those opportunities. This announcement marks a defining moment for agentic commerce – similar to the early days of the web, when foundational protocols like HTTP quietly enabled scale – signaling that the infrastructure layer of the agentic age is now beginning to take shape. Although this direction may not come as a surprise, the “how” is important to consider; that is, how exactly is Google planning to deliver agentic commerce at scale for its extensive network of retail partners?
In the announcement, Google outlines several solutions, including Universal Commerce Protocol (UCP), Business Agent, and Direct Offers. While Direct Offers largely function as an out-of-the-box incentive-based enhancement to existing Google Ads campaigns, both UCP and Business Agent focus on deeper, more meaningful integration with a retailer’s commerce infrastructure.
Universal Commerce Protocol (UCP)
You would be right to ask, ‘Didn’t OpenAI already introduce the Universal Commerce Protocol?’ They launched the Agentic Commerce Protocol (ACP) in September 2025. However, despite the similar naming conventions, the two protocols differ both technically and philosophically.
The key distinction lies in their approach to integration. OpenAI places a strong emphasis on strict standardization, whereas Google prioritizes flexibility within a standardized framework. While the purpose of any protocol is standardization, Google is focused on standardizing how platforms communicate, and not the result. This means that with Google’s approach to UCP, retailers retain greater control over how integrations are customized and how the consumer experience unfolds. For example, Google has confirmed plans to support opt-in for marketing communications within agentic flows – a capability that ACP does not currently enable. In addition, UCP will support both native checkout and embedded checkout. Embedded checkout allows brands with complex, regulated, or highly customized purchase journeys to integrate their own web checkout directly into Google’s workflow, whereas ACP today only supports native checkout.
It’s important to note that UCP and ACP are currently the only mechanisms through which retailers can become eligible for agentic checkout, making protocol adoption a prerequisite – not an enhancement – for participation.
Business Agent
With Business Agent, Google is bringing agentic, conversational experiences into standard Search Ads. While traditional search is compressing and declining – largely due to the rise of AI-driven chat interfaces – there remains a sizable audience that has not yet made that transition. Embedding agentic experiences within familiar search environments lowers the barrier to adoption. It introduces conversational AI where users already feel comfortable, rather than forcing behavioral shifts. Again, customization is central to Google’s approach. Retailers can train a Business Agent using their own data, ensuring that responses are not only accurate but also aligned with the brand, and its product positioning.
Google Merchant Center (GMC)
Perhaps the most significant development is the expanded role of Google Merchant Center (GMC). GMC is evolving into a core infrastructure layer for agentic commerce, ensuring products remain discoverable throughout the customer journey.
Business Agent will live within GMC, and will learn to answer product-related questions directly from a retailer’s product catalog. More notably, Google announced the rollout of dozens of new data attributes within GMC, specifically designed to optimize visibility within conversational AI environments such as Gemini and AI Mode.
These attributes are notable not just because they are new, but because they are not currently supported within OpenAI’s feed schema. They allow agents to understand how products relate to one another, rather than simply describing individual characteristics. Importantly, attributes such as faq also enable brands to speak to multiple customer personas from a single feed – a role historically played by the website or user navigation, but now increasingly mediated by the agent itself.
Examples of these new data attributes include:
- Product_faq, which allows retailers to upload product-specific questions and answers.
- Product_use_cases, which defines the specific scenarios the product is best suited for.
- Compatible_accessories, which offers a list of IDs for products that work with the main item.
- Product_substitutes, which suggests alternatives if the main item is out of stock, allowing the agent to save the sale.
- Native_commerce, which is a toggle that tells Google if the product is eligible for Agentic Checkout.
- Checkout_eligibility, which defines if the item can be purchased by a “Human Not Present.”
Together, these attributes provide conversational AI systems with richer context, enabling more relevant and informed interactions. However, they also introduce additional operational complexity – particularly for retailers managing large or frequently changing product catalogs.
What This Means for Retailers
It’s clear that Google Merchant Center is becoming the gateway to product visibility in AI-driven discovery. Investing in a robust, well-structured GMC setup is no longer optional – it’s foundational.
Attributes like product_use_cases, in particular, create new opportunities for differentiation. Strong product titles and descriptions will always matter, but the competitive landscape has now significantly expanded. Historically, most feed attributes beyond titles and images functioned as hygiene factors – focused on compliance and incremental gains, with largely one “correct” configuration. Today, those same attributes have become areas for active optimization, where multiple strategic approaches are possible and the upside increasingly favors brands willing to test, learn, and refine.
Our Perspective
This is a significant shift in how customers discover, evaluate, and purchase products. When implemented strategically, agentic commerce will empower brands to lead – not follow – and ultimately outperform.