El agentic tradeThat is, purchases managed by artificial intelligence agents acting on behalf of the user have gone from being a futuristic idea to becoming the major focus of large technology companies for the coming years. Google, in particular, identifies it as one of its strategic priorities for transforming advertising, e-commerce, and the relationship between brands, creators, and consumers.
According to the plans that the company has been detailing in its latest communications, 2026 is shaping up to be the year in which this AI-assisted purchasing model will take off. The transition from testing and pilot programs to a much broader rollout will be key. The key will be combining new payment infrastructures, advanced capabilities of models like Gemini 3, and standards specifically designed to enable agents to make decisions and execute transactions securely.
What is agent trading and why is it gaining so much traction?
When we talk about agent trading, we are referring to an environment in which Artificial intelligence agents operate directly within the purchasing processThey no longer just recommend products or show ads: they can interpret needs, compare alternatives, choose an option and complete the payment following rules set by the user or the company.
This idea is supported by the maturity of the so-called Agentic AIThese systems are capable of acting with a degree of autonomy on behalf of third parties. Instead of the consumer having to spend time reviewing reviews, product information, or shipping conditions, they define their preferences—maximum price, favorite brands, delivery times, sustainability criteria—and the agent takes care of the heavy lifting.
Several international analyses suggest that this change could be profound. Some reports estimate that Up to 30% of the value of global e-commerce could be influenced by AI agents in the next decadeThis would involve managing trillions of dollars in automated decisions and purchases. It's not just about improving the user experience, but about redesigning how transactions are planned, executed, and settled.
In this model, the consumer's role shifts: from manually performing each step to becoming the one who sets the limits and objectives that the agent must meet. The relationship between customers, businesses, and technology platforms becomes more indirect, but also more fluid and personalized, with automation at its core.
Google's bet: from conversational search to AI Mode
Google's strategy for this new scenario revolves around the transformation of its search engine and advertising environments. The company describes how the AI is changing the way people search, compare, and buyleaving behind the classic model based solely on keywords to make way for conversational queries, combinations of text and images, and much more complex requests, as well as features such as shopping, orders and shipping tab.
In that context, the so-called AI modeA search experience where results are presented in a more conversational and contextual way. Within this mode, Google is testing advertising formats integrated into the AI experience itselfnot only as sponsored links on the side, but as product recommendations that appear alongside organic responses, always clearly marked as promoted content.
This approach is being tested with retail businesses and in verticals such as travelwhere planning is often complex. The idea is that the user can submit a broad request to the system—for example, organizing a getaway with a specific budget and dates—and that the agent, in addition to preparing proposals, can display relevant offers from brands that fit that scenario.
In parallel, Google is promoting formats such as Direct OffersThese are designed so that companies can present personalized discounts or promotions to users who are close to completing a purchase, without having to modify their general offer for other visitors. In this way, the ad is integrated into a broader conversation, instead of appearing as an isolated impression.
Gemini 3: the engine of the next generation of commercial experiences
Much of this transformation relies on Gemini 3, the AI model that Google presents as the most advanced in its catalog for reasoning and context understanding tasks, as demonstrated its arrival at GmailThis system is already integrated into their advertising tools, with the intention of improving the ability to understand the intent behind each query and to generate content tailored to each campaign.
In the environment of Google Ads Asset StudioGemini 3 powers utilities like Nano Banana and Veo 3, designed to produce creative and audiovisual assets faster and at a lower cost for advertisers. It also forms the basis of solutions such as AI Max, which expand the reach of search campaigns to include new queries without requiring brands to manually define all combinations of terms.
Internal data that Google has been sharing points to an accelerated growth in this automated use: By 2025, the volume of creative assets generated with Gemini would have tripled.And in the last quarter of the year alone, the figure of almost 70 million pieces produced for campaigns in AI Max and Performance Max would have been reached.
This ability to generate content at scale aims not only to improve the efficiency of marketing teams but also to fuel the agent commerce ecosystem. The more accurate the understanding of what a person wants and the more tailored the marketing messages, the easier it will be for AI agents to accurately select products, services, or combinations that fit each specific case.
From the creator economy to measurable commercial impact
Another pillar of Google's strategy involves leveraging the growing importance of creator economyespecially on platforms like YouTube. The company believes that content creators have become key players in building trust, trends, and influence within their communities, something that aligns perfectly with the logic of agent commerce.
The idea is to use artificial intelligence to to analyze in detail both the content and the audiences of each channelGoogle identifies which communities are most aligned with specific brands or products. Based on this information, Google aims to connect advertisers and creators almost instantly, matching their audiences with the campaign's business objectives.
As the company itself has explained, this approach allows for the transformation of the organic influence of creators on a more direct and measurable commercial impactFor companies, it represents a way to reach very specific niches with tailored proposals, while for creators it opens the door to collaborative models in which recommendations are integrated more naturally into the content.
In an advertising environment where brands demand a higher return on investment and more accurate measurement, these AI-based tools aim to provide more consistent data on what content generates real sales, which segments respond best, and how to optimize creativity based on audience response.
AP2 and UCP: the protocols that make agent trading possible
The ambition for AI agents to be able to buy on behalf of the user requires more than advanced linguistic models: it requires dedicated infrastructure for payments, identity and security, including advances in fintech and online bankingIn this area, Google has been introducing two key pieces: the Agent Payment Protocol (AP2) and the Universal Trade Protocol (UCP).
AP2 is designed to allow agents to initiate and complete payment transactions on behalf of individuals or businessesWhile respecting security and regulatory compliance frameworks, UCP is envisioned as a standard for connecting merchants, payment partners, and AI agents throughout the entire purchase journey, from digital user identification to order completion.
These protocols are already being used in the United States in an initial practical deployment. UCP payment process It allows those searching for products in AI Search Mode or in the Gemini app Buy items directly from platforms like Etsy and Wayfairwithout having to leave the conversational environment. This functionality is also expected to be extended to retailers such as Shopify, Target, and Walmart.
The company assures that hundreds of technology companies, payment providers, and retailers They have shown interest in integrating into this standard. Beyond the initial cases in retail, the intention is that, over time, this interoperable infrastructure can be applied to other sectors, from services to travel or subscriptions, always with AI agents acting as intermediaries.
Privacy, security and trust: essential conditions for scaling
As AI takes on a more active role on behalf of the consumer, logical questions arise about privacy, data protection and control by the user. Google insists that the deployment of these agents is based on the same security principles that have guided its products in recent decades, with special attention to transparency and regulatory compliance, and to offer Safety tips for your shopping.
In its public communications, the company's management emphasizes that “Empowering agents to act on behalf of consumers and businesses” Strict standards are maintained to ensure that the fast option remains the safe option. This includes controls over who can authorize payments, how the identity of the agent is verified, and what traceability exists for the decisions made.
At the same time, the rise of agent trading is forcing an adaptation of the payments infrastructure and financial systems This leads to a scenario where it is not always a person who initiates the operation, but rather an autonomous system. Among the challenges are the identification of these agents, interoperability between platforms, and the ability to process transactions in real time without compromising security.
In Europe, where regulations on data and payment services are particularly stringent, these types of proposals will have to comply with rules such as GDPR or the PSD2 regulation and its evolution. How these issues are resolved will be crucial for agent trading to expand on a large scale in European markets as well.
How should businesses and advertisers adapt to the agent era?
For companies, the advancement of agent trading not only means taking advantage of new advertising channels, but rethink how their catalogs are structuredsystems and processes to be able to interact effectively with AI agents. Reports on this topic agree on several basic requirements.
First, it is necessary to have structured and accessible product dataThis allows agents to accurately understand what's being offered, the terms and conditions, and the differences between options. Incomplete, outdated, or unclear information makes it difficult for AI to make sound decisions on behalf of the user.
Secondly, it is recommended to have updated information in real time Regarding prices, stock levels, delivery times, or restrictions, agents perform best when working with reliable data, and any discrepancies can lead to purchasing errors, returns, or a loss of trust.
Furthermore, agent trading favors models based on Open APIs and architecturesThis will allow for more direct integration with payment platforms, recommendation engines, and smart assistants. For many retailers and brands, this will mean reviewing legacy systems and moving toward more modular and connectable infrastructures.
Finally, internal processes—from order management to after-sales service—will have to adapt to a higher degree of automationCustomer experience is no longer limited to the human interface, but to how different intelligent systems interact with each other on behalf of users, which forces a review of metrics, workflows and quality criteria.
Building on these developments, agent-based commerce is emerging as a significant evolution in the digital ecosystem: agents who interpret needs, explore the market, negotiate offers, and close purchases without requiring user intervention at every click. This is supported by models like Gemini 3, standards such as AP2 and UCP, and conversational environments like AI Search Mode and the Gemini app. There is still a long way to go—especially in markets like Europe, with stricter regulations—but the moves by Google and other major players indicate that the race toward increasingly assisted and automated shopping is already underway.
