

Agentic commerce is a shopping model in which artificial intelligence can take actions on a person’s behalf, within the limits of their authorization. Depending on the capabilities available, an agent can clarify a need, compare products based on a budget or intended use, then prepare or initiate a purchase. Its level of autonomy depends on the shopping journey and the approvals required.
Announcements made in 2026 are bringing this shift into practice. On November 4, Google announced the gradual rollout of direct checkout in Canada through AI Mode in Google Search and the Gemini app. Simons, lululemon, and select Shopify merchants were among the eligible retailers named. (1)
It can reduce friction in complex shopping carts, large catalogs, or purchases that require extensive comparisons. To take advantage of this opportunity, a catalog must make it possible to find the right item, compare its features and variants, and check availability and delivery or pickup options. Clear, current information helps AI systems recommend products that match the needs expressed, while connected systems confirm the actual purchase conditions. Preparation therefore involves both making products discoverable and ensuring orders can be fulfilled.
How AI is transforming Black Friday 2026
Read more
The Universal Commerce Protocol (UCP) is an open standard launched in January 2026 by Google alongside commerce companies, including Shopify. It establishes a common language for interactions between agents, merchant systems, and payment providers, from product discovery through post-purchase support. (2)
In Canada, Google plans to begin expanding access in 2027 through a simplified UCP integration in Merchant Center. For retailers, the priority is to check their eligibility and the quality of the data they submit. (1)
Access the on-demand webinar to see the examples and applications presented using Shopify
Register for the webinarAn agent must be able to distinguish between variants, understand how a product is used, and access its terms of sale. Prices, inventory, and lead times must also reflect what the business can actually deliver. In the Canadian experience announced by Google, the retailer remains the merchant of record. (1)
These requirements align with the foundations of unified commerce. A product information management (PIM) system can centralize product attributes and content, while enterprise resource planning (ERP) and order management (OMS) systems supply commercial and operational data. Responsibilities vary by architecture, making it important to define a source of truth for each data type and ensure that information reaches every sales channel.

Agentic commerce will only be as good as the product content behind it. If an AI system cannot understand what a product is, who it is for, how it compares, or whether the information is reliable, it cannot confidently recommend it.
That is why the product data foundation matters so much. Commerce teams need structured, complete, and consistent product information across titles, attributes, variants, descriptions, media, feeds, and channel requirements.
As a Novatize partner, Plytix provides a suite of tools like PIM, DAM, Feed Management, AI enrichment, Product Data Sheets, and Brand Portals into unified solution that helps ecommerce teams manage and optimize product content that actually performs on their storefronts, marketplaces, retailer channels, search engines, and emerging AI-driven experiences.
At Plytix, we see product data as much more than an operational requirement. We see it as a growth asset. PIM is our flagship product, but Plytix is not just a PIM. We’re building a multi-product platform for content-led growth, helping commerce teams turn product information into better discovery, stronger PDPs, better channel performance, lower returns, and more scalable growth.
— Morten Poulsen, Founder & CEO @Plytix
Shopify co-developed UCP with Google around real commerce requirements, including payment rules, discounts, and delivery options. This involvement helps align the platform with the interactions the protocol needs to support. Shopify businesses therefore have a relevant starting point for evaluating these channels. They still need to confirm which capabilities are available in their market and whether their specific business rules are supported. Custom pricing or an approval process requires a closer assessment than a standard purchase.
“We’ve moved from a web designed for our eyes to a web built for agents. With structured data and a unified architecture, Shopify enables merchants to be immediately discovered and understood by artificial intelligence.”
— Mathilde Vandenbosch, Principal Partner Solutions Engineer, Shopify
Shopify Winter ’26—AI, agentic commerce, and B2B features our team is watching
Read moreConversational tools can help people express a need and understand an offering. An agentic shopping journey adds the ability to take action based on that conversation. Both approaches can coexist within the same experience.
Announcements made in 2026 marked progress in standardizing these interactions and bringing them to new markets. For merchants, it is essential to distinguish between what a protocol supports and what is actually available within a specific interface.
The 2025 announcement involving ChatGPT already represented a step toward agentic commerce, notably through the introduction of capabilities for making purchases directly within a conversational interface using the Agentic Commerce Protocol (ACP), developed by OpenAI and Stripe.
The 2026 Universal Commerce Protocol (UCP) announcement builds on this evolution, aiming to standardize interactions between AI agents, commerce platforms, and transaction systems throughout the shopping journey. These initiatives reflect the same shift: merchants now need to make their offerings discoverable and understandable, as well as accessible to systems capable of coordinating transactions.
For more context, read our analysis of the 2025 announcement.
Simons and lululemon illustrate the arrival of this channel in Canadian retail. Google names them among the retailers eligible for direct checkout, without detailing how transactions work at either company. The following hypothetical scenario illustrates this type of shopping journey and the data needed to support it.
Someone is looking for travel pants in a specific size, within a set budget, with delivery before their departure. An assistant could compare options, check whether the right variant is available, and prepare the purchase, depending on the channel’s capabilities. For the merchant, useful attributes go beyond the product name and color. Fit, fabric composition, a size guide, and return policies help the customer choose. Inventory for the specific variant and available delivery options then determine whether the purchase can proceed.
This is why product content quality directly connects to content-led growth. Better product data does not only make internal workflows cleaner. It helps products get found, understood, recommended, and purchased with more confidence.
— Morten Poulsen, Founder & CEO @Plytix
To take advantage of the opportunities agentic commerce creates, start by assessing your shopping journeys and data.
Which products are difficult to understand? What information is missing when choosing a variant? Where do pricing or inventory discrepancies occur? This assessment helps prioritize fixes based on their impact on the purchase.
Next, work on the integrations between your eCommerce platform and the systems that support your offering. Test a limited journey using real scenarios: an unavailable product, an address outside your delivery area, an incompatible promotion, or an order requiring approval. This will help you determine what the agent can handle and when human intervention is needed.
Governance should define permitted actions, required confirmations, and how cancellations or substitutions are handled. To measure the project’s value, track completed purchases, errors, returns, and margins alongside visibility in AI interfaces.
Is your offering easy for AI to understand and compare? Explore eight prompts to start assessing your readiness and identify improvements to your catalog and content.
Assess your agentic commerce readiness
Learn moreThe ability to provide actionable information can influence how an offering appears in an assisted shopping journey. Missing information about a variant or delivery option can prevent a purchase from moving forward, even when the product meets the customer’s needs. Operational reliability therefore becomes a factor in customer acquisition strategy.
Statistics shared by Shopify at DotDev 2026 illustrate this channel’s potential. Over one year, traffic from AI assistants grew more than eightfold, while orders generated by these assistants increased nearly thirteenfold. More than half of these visitors land directly on a product page. Their conversion rate is nearly 50% higher than that of organic search traffic, and Shopify also reports an average order value approximately 14% higher for these visitors.
These metrics concern traffic from AI assistants; they do not mean transactions are completed entirely within those interfaces. They reinforce the importance of product pages, which can become the first point of contact with customers who have already clarified what they need.
For 2027, the opportunity is to test shopping journeys that are relevant to your customers and measure their contribution to sales and profitability.
Explore the key takeaways from Shopify DotDev 2026
In an agentic commerce environment, SEO remains essential for capturing demand from people searching online, but it is no longer sufficient on its own. As AI agents become shopping intermediaries capable of comparing products, making recommendations, and even executing transactions, businesses also need to help language models discover and understand their offerings. This is the purpose of Generative Engine Optimization (GEO), also referred to as Generative Search Optimization (GSO): structuring information, product data, and evidence of credibility to support accurate interpretation, citation, and selection by AI systems. The objective now extends beyond a position on a search results page to being selected as a reliable, actionable answer in shopping journeys where AI helps guide the decision. In this context, GEO can provide a competitive advantage. It improves machine readability, the consistency of information across products, prices, availability, and policies, and the ability to appear in a new acquisition and conversion channel shaped by customer intent.
For a closer look, read our article: SEO vs. GSO: Which strategies can help grow your eCommerce business with AI?
In B2B, agentic commerce could accelerate a transformation already underway: automating recurring purchases, streamlining complex buying journeys, and coordinating advanced business rules. While B2C primarily benefits from smoother shopping experiences, B2B businesses could see measurable productivity gains, particularly when dealing with large catalogs, product configurations, negotiated pricing tiers, compliance requirements, specific delivery terms, or internal approval policies. An agent could, for example, help a buyer turn a purchasing need into a cart that complies with contract terms, including pricing, availability, permitted substitutions, lead times, and minimum order quantities (MOQs), then initiate the transaction according to the company’s authorization rules. For B2B organizations, the priority is therefore strategic: structure data, clarify rules, and improve the reliability of operational information to prepare for working with agents, both to sell and to reduce purchasing friction while improving service to business accounts.
Read our strategic guide to overcoming complex B2B eCommerce challenges on Shopify

Would you like to discuss this with a Novatize specialist?

Pierre-Olivier Brassard
Sources





