2026 marks the beginning of the agentic commerce era.
At the start of the year, Google released the Universal Commerce Protocol (UCP) in collaboration with PayPal, Visa, and other payment processors.
Additionally, online retail giants like Zalando, The Home Depot, Walmart, and Shopify are among the protocol’s co-developers and early adopters.
UCP is currently viewed as one of the foundational building blocks of agentic commerce, as it ensures the safety of transactions AI agents complete on behalf of shoppers.
However, the protocol is currently only available in the US as Google is still in the process of onboarding merchants. Its rollout is expected to continue over the next couple of years, removing the biggest obstacle to worldwide adoption of agentic commerce.
Let’s explore how agentic commerce will affect the ecommerce industry and change online shopping in the coming years.
Key takeaways
- Agentic commerce could generate up to $5 trillion in revenue worldwide by the end of the decade.
- 15% to 25% of ecommerce sales in the US could be completed through agentic commerce by 2030.
- Approximately half of US online shoppers are expected to actively use AI agents to find and purchase products over the next five years.
- Less than 1% of ecommerce businesses use agentic commerce in 2026, but this number is expected to increase to 33% by 2028.
- 47% of consumers say they are willing to use AI to complete repetitive purchases.
Defining agentic commerce
Agentic commerce refers to a fully automated online shopping process in which AI agents handle all steps, from product research to price comparisons and checkout.
Rather than just finding products that match a shopper’s criteria, AI agents are now capable of going a step further and completing a purchase.
For instance, if a consumer is searching for a new pair of Nike sneakers under $75, an AI agent can find listings that meet the criteria, compare prices across models, and purchase the product with minimal human oversight.
Fully autonomous agents are still novel, as the infrastructure required for their safe operation is under development, which is why their availability remains limited.
However, early data suggest that agentic commerce will become the dominant form of online shopping over the next 10 years.
The statistics show that in 2026, a significant portion of consumers who frequently buy products online are already using AI tools to research products, find the most affordable options, and add items to shopping carts.
Nonetheless, 51% of consumers remain uncomfortable with AI agents storing and using their credit or debit card details, which is why the adoption rate of autonomous AI spending remains low.
How does agentic commerce work?
The nature of agentic commerce is conversational.
A user simply enters a prompt containing information about a product or service they’re looking for. AI agents are trained to analyze the request, contextualize it, and offer a suggestion for moving forward in the buying process.
For example, if a user enters a broad natural language prompt, stating they are looking for a new pair of pants, the agent will ask the user to provide more information about the size, color, material, brand, and price range before searching the internet for product listings that match the criteria and displaying them in the same thread.
Agentic commerce tools are capable of making decisions independently due to their impressive reasoning capabilities. They can break complex requests into actionable steps, determine which tool to use for each step, and achieve their goal.
AI agents never actually visit a retailer’s site. Instead, they send API calls to the site’s backend using UCP and similar protocols to request structured data, which is then processed and displayed in the thread a user sees.
After the user approves the purchase, the AI agent places the order through the same protocol it used to find the recommended product.
Understanding the agentic commerce protocol layers
Protocol layers are the language of agentic commerce.
Aside from ensuring the safety of the entire buying process, protocols transform the information AI agents find on the internet into readable data.
They establish standardized and universally understood commands for product discovery, cart creation, and finalizing the purchase.
Currently, OpenAI’s Agentic Commerce Protocol (ACP) and Google’s Universal Commerce Protocol (UCP) are the best-known examples of front-facing conversational protocol layers.
Anthropic’s Model Context Protocol (MCP) enables an AI agent to connect to different data sources and API endpoints. Both ACP and UCP use MCP binding to look inside merchants’ servers and request information about specific products.
Additionally, Google has developed an Agent-to-Agent (A2A) protocol that enables different AI tools to communicate with one another. This protocol allows Gemini, Perplexity, ChatGPT, and other AI platforms to communicate with commerce agents on a retailer’s site or an online marketplace.
Amazon, eBay, and other online marketplaces are taking a walled-garden approach, preventing the A2A protocol from interacting with their native AI agents to protect their user data.
It’s also worth noting that AI agents using the currently available protocol layers cannot complete purchases without prior human authorization.
At the moment, most protocols require user approval before initiating checkout, so an AI agent cannot complete a purchase without human authorization.
However, Visa, MasterCard, and American Express developed Agent Payment Protocols (AP2) that use cryptographic spending mandates to enable users to grant AI models the necessary permissions upfront and make repeated purchases.
The current state of AI agents for commerce
As noted earlier, the availability of fully autonomous AI agents capable of completing purchases on the user’s behalf with prior human approval is still limited.
Perplexity Pro subscribers in the US have access to the ‘Buy with Pro’ feature that lets them complete purchases without leaving the platform.
Similarly, Gemini users in the US can use its agentic commerce features to purchase products from Etsy, Wayfair, and other eligible US-based retailers.
Google is currently in the business onboarding stage, and by the end of 2026, consumers will be able to use Gemini’s native checkout option to buy products from Walmart or Target.
Neither of these shopping tools is available worldwide, as complex international payment regulations, shipping restrictions, and a lack of structured product data curtail their global rollout.
Still, Gemini’s native checkout feature should become available to shoppers in the EU and the UK in early 2027, while consumers in Asia and Latin America will have to wait until at least 2028 to start buying products online with Gemini.
What’s the difference between agentic AI and agentic commerce
Agentic AI is an autonomous AI system trained to understand complex goals and create plans to achieve them with minimal or no human oversight.
Agentic commerce is an application of agentic AI in ecommerce designed to solve problems related to product discovery, pricing comparisons, negotiations, or fulfillment.
However, both types of artificial intelligence systems have the following traits:
- Autonomy – Agents receive human input and work independently to solve the problem.
- Objective achievement – Agentic AI models aren’t focused on solving a single task, such as finding the most affordable product within a specific range, but rather on achieving a goal, whether it is completing a purchase or creating code for a new mobile app.
- Context interpretation – Agents extract data from multiple sources to understand the context and provide users with optimal solutions.
- Adaptability and learning – Agentic AI systems leverage a user’s behavioral data, such as past purchases, to learn their preferences and provide solutions that better align with their criteria.
In other words, agentic AI is the underlying technology that powers agentic commerce by enabling agents to recommend products to online shoppers and complete purchases on their behalf.
AI shopping assistants and the future of online marketplaces
The world’s largest online marketplaces, including Alibaba, Amazon, and eBay, are heavily investing in the development of AI assistants and platform-native AI-powered algorithms.
In May 2026, Amazon released Alexa for Shopping, which replaced Rufus and Alexa+.
Amazon’s new AI agent can buy products for shoppers within the platform if given permission, as well as execute purchases of products that are out of stock on Amazon using the ‘Buy for Me’ feature from third-party websites.
Furthermore, Alibaba’s Qwen app will be integrated with Taobao and Tmall marketplaces by the end of 2026, enabling millions of online shoppers to fully automate the buying process on these marketplaces.
On the other hand, eBay has recently taken a strong stance against Buy-for-Me bots, banning their use across its marketplaces.
The ecommerce giant is actively developing its own in-house AI shopping assistant capable of automatically completing the checkout process.
So rather than embracing the UCP and other available protocols, some of the world’s largest marketplaces are taking the walled-garden approach and are focusing on developing agentic commerce solutions within their own ecosystems.
Getting your eBay listings ready for AI agent interactions
Although it is still in its infancy, agentic commerce is poised to transform the ecommerce industry and become a dominant form of online shopping by the end of the decade.
The change is already underway, as online marketplaces have already implemented tools that change how shoppers discover products and affect the visibility of product listings.
That’s why eBay sellers should optimize their listings for generative engines and ensure that their stores are ready for the agentic commerce era.
Would you like to improve your performance across eBay marketplaces with Generative Engine Optimization? Schedule a call to learn more about how Webinterpret can get your product listings ready for the agentic commerce era!
About Webinterpret
Webinterpret supports merchants selling on eBay.
Our AI-based solutions enable more effective selling through automated listing localisation, optimisation, advertising, and returns, and ensure all products placed on EU markets are GPSR-compliant.
By giving your international and domestic customers a full, end-to-end local shopping experience, Webinterpret improves your conversion rates and helps establish your business locally and globally.
