Google Unveils AP2: A New Protocol for AI Agent-Driven Purchases

Google launches new protocol for agent-driven purchases | TechCrunch

On Tuesday, Google announced a new open protocol for purchases initiated by AI agents -- automated software programs that can shop and make decisions on behalf of users -- with backing from more than 60 merchants and financial institutions. Called the Agent Payments Protocol (AP2), the system is meant to be interoperable between AI platforms, payment systems and vendors, providing a traceable paper trail for each transaction. In a post announcing the protocol, Google executives emphasized their commitment to openness. "We are committed to evolving this protocol in an open, collaborative process, including through standards bodies, and invite the entire payments and technology community to build this future with us," wrote Stavan Parikh and Rao Surapaneni, who are VPs at Google and Google Cloud, respectively. The full specification for AP2 was posted to GitHub in conjunction with the announcement. The protocol is built for a future in which AI agents routinely shop for products on customers' behalf, and engage in complex real-time interactions with retailers' AI agents. One example in Google's post imagines a chatbot user asking their agent to shop for a bike trip, which triggers a spontaneous time-sensitive bundle offer from a bike shop's agent. In another example, a user asks for travel and lodging for a weekend vacation, giving only the dates, location and budget. "The agent can then interact with both airline and hotel agents, as well as online travel agencies and booking platforms," the post explains, "and once it finds a combination that fits the budget, it can execute both cryptographically-signed bookings simultaneously." Enabling that kind of transaction is complex, from both a technological and social standpoint. AP2 requires agents to register two separate approvals before a purchase can be made: first the "intent mandate" (essentially telling the AI, "I'm looking for a polka dot neck tie"), which enables the agent to search for a specific item and negotiate with sellers; then the "cart mandate," which gives final approval for a purchase once a specific item has been found. The protocol also contains a provision for fully automated purchases, in which the agent is permitted to automatically generate a cart mandate once an item is found. Those circumstances require a more detailed intent mandate, specifying price limits, timing, and other rules of engagement. In either case, the goal is to maintain an auditable trail that can be re-examined in cases of fraud. In collaboration with cryptocurrency outfits Coinbase, Metamask and the Ethereum foundation, Google also produced an extension that would integrate the cryptocurrency-oriented x402 protocol, allowing for AI-driven purchasing from crypto wallets. A number of other tech companies are working on their own agentic purchasing systems -- most notably Perplexity, which allows for a Buy With Pro service in its agentic browser. The payment provider Stripe also produces software tools for agentic purchasing on its platform, though they are not as comprehensive as AP2. Like any protocol, the impact of AP2 will depend on its support from other players in the ecosystem -- most notably, developers building agentic purchasing systems. But AP2 has already won the support of major financial providers like Mastercard, American Express and PayPal, giving the protocol a significant immediate footprint.

Google's new open protocol secures AI agent transactions - and 60 companies already support it

The protocol is launched with the support of more than 60 organizations. The next level of AI assistance in e-commerce is agent-led AI transactions, where your agent can place orders on your behalf, saving you time as a consumer and connecting your product to buyers if you're a merchant. Also: Gartner says add AI agents ASAP - or else. Oh, and they're also overhyped While the technology is promising, there are also security concerns because an agent has access to your payment details. To address those concerns, Google has launched its new Agent Payments Protocol (AP2), which the company describes as an "open protocol developed with leading payments and technology companies to securely initiate and transact agent-led payments across platforms." AP2 can be used as an extension of the Agent2Agent (A2A) protocol and the Model Context Protocol (MCP). Similar to how MCP allows users to connect AI tools to their databases with confidence, the AP2 standard lets users, merchants, and payment providers transact with AI agents securely. Also: How Cisco plans to stop rogue AI agent attacks inside your network For example, Google said AP2, which helps verify that a user authorized an agent to act, validates to the merchant that the agent is performing the way the user requested, and, most importantly, helps identify accountability should the action not go according to plan, which is one of the biggest concerns when it comes to agents. Also: 5 ways to be a great AI agent manager, according to business leaders AP2 supports various payment types, including credit cards, stablecoins, and real-time bank transfers, which serve as a single standard that agents and merchants can rely on for all transactions. AP2 uses Mandates, which are tamper-proof, cryptographically signed digital contracts that contain proof of a user's intentions. Specifically, there is an Intent Mandate, where the user requests that the agent either find a product for them or purchase it on their behalf, and then a Cart Mandate, which allows the agent to carry out the transaction. Also: A whopping 30% of business leaders are comfortable taking orders from AI agents - are you? In real-time purchases with a user present, after issuing an Intent Mandate to look for the product, the approval of the purchase is signed as a Cart Mandate. Also: Your next job interviewer could be an AI agent - here's why that's a good thing In delegated tasks, where a human isn't present, the human signs a detailed Intent Mandate upfront, with the specific rules of the engagement, such as price limits, which pre-authorize the agent to generate a Cart Mandate when the conditions are met. In either case, the Cart Mandate involves the user securely linking their payment details. The AP2 launched with 60 organizations, including Accenture, Adobe, American Express, Coinbase, Confluent, Mastercard, Okta, PayPal, Salesforce, 1Password, and many more. Also: What are AI agents? How to access a team of personalized assistants For other organizations that want to get started, Google has made the technical specifications, documentation, and reference implementations available on its public GitHub repository, which will be updated regularly with additional reference implementations from the tech company.

Google unveils masterplan for letting AI shop on your behalf

Mastercard, American Express, Coinbase, and PayPal sign up at launch Google has given the go-ahead to a plan that lets AI agents make purchases on your behalf and, on Tuesday, released its Agent Payments Protocol (AP2) to make it happen. The system comes with touted safeguards that are intended to prevent thieves from draining bank accounts. The principle is that shoppers can use AI agents to create a shopping list, exchange information with merchants, and complete payment transactions, without the need for final, human approval. For example, a music fan could tell an agent to buy concert tickets that go on sale at midnight and then go to sleep, knowing that the agent would buy the number and location of tickets they had asked for (presumably with a price limit). Youtube Video For vendors and payment processors, the protocol provides a verifiable paper trail - or mandate - that they can back trace and check. End users can set up an Intent Mandate, which says what they want and how much they wish to pay for it. If the purchase is taking place in real-time (ex: "find me a pair of white shoes"), then the human can personally approve the purchase by signing a Cart Mandate. If the purchase takes place when the human is offline (ex: "get me two tickets at midnight and don't pay more than $100 each"), the Intent Mandate can also give permission for the agent to generate its own Cart Mandate. "This chain of evidence culminates in securely linking your payment method to the verified contents of the Cart Mandate," Google wrote in a blog post. "This complete sequence -- from intent, to cart, to payment -- creates a non-repudiable audit trail that answers the critical questions of authorization and authenticity, providing a clear foundation for accountability." We've asked Google for more details on security checks that can be built into the system and will update this story if we hear back. The AP2 scheme is an open protocol that works with Google's existing Agent2Agent (A2A) data-sharing systems and Anthropic's Model Context Protocol (MCP) to allow secure information flows. It's part of a wider move to allow AI engines to communicate securely with each other, something the Chocolate Factory has been pushing hard of late. At launch, Google has signed up over 60 companies with major players like Mastercard, PayPal, American Express, and Worldpay getting on board. Salesforce, Red Hat, Adobe, Intuit, and Cloudflare have also joined. AP2 is also supporting cryptocurrency payments using the x402 protocol, to allow digicash transactions using the same security system. Coinbase, Metamask, and the Ethereum Foundation have already signed up. "x402 and AP2 show that agent-to-agent payments aren't just an experiment anymore, they're becoming part of how developers actually build," said Erik Reppel, head of engineering at the Coinbase Developer Platform in a canned statement. "Bringing x402 into AP2 to power stablecoin payments made sense - it's a natural playground for agents to start transacting with each other and testing out crypto rails. And it's exciting to see the idea of agents paying each other resonate with the broader AI community." While the system has some big-name backers, security remains critical. Given the propensity of online thieves to abuse payment processing, hopefully Google will have adequate protections in place. Otherwise, you may be seeing headlines about AI-driven theft in the future. ®

Google's new Agent Payments Protocol (AP2) allows AI agents to complete purchases -- is your enterprise ready?

In an ideal world, AI agents would complete all the tasks we ask of them. With the rise of browser agents that can surf the web and click around on interactive menus and elements like humans do, the possibility of that future world is tantalizingly close. While the current crop of AI agents can browse the internet for flights or help enterprises compare potential product providers, they still typically stop short of going through with a purchase or transaction -- certainly not without getting a user's express, manual approval. Now Google, along with 60 other major organizations is introducing a new, open source payment protocol any enterprise or developer can adopt -- with the goal of making AI agentic payments more convenient, more secure and traceable, and ultimately, more seamless, letting an AI agent actually complete a purchase even if the human user isn't there right at the transaction time to bless it. This should also unlock new ways for people to buy things online, such as having a human user instruct an agent to find deals and buy when an item reaches a certain low price, for example. Announced today, Agents Payments Protocol, or AP2, is aimed at establishing a common ground for companies within a digital purchase ecosystem, including merchants, banks, agents, model providers, as well as customer relationship management (CRM) and accounting platforms, to determine whether a user authorizes a transaction through an AI system, and who can be held accountable for errors. Some of the collaborators include American Express, Ant International, Airwallex, Cider, Coinbase, Etsy, Intuit, JCB, Lazada, Mastercard, PayPal, Revolut, Salesforce, ServiceNow, UnionPay International, Worldpay and Zalora. A standard would provide the necessary guarantee that requests to pay for something are legitimate and offer guidelines if something does go wrong. Even with this concerted effort from the search-turned-AI giant, it will likely take some time before consumers and enterprises are comfortable with humans being barely involved in transactions. As of today, consumers cannot use AP2-based payments today in any publicly known, real-world product or service. The protocol is ready for developers to explore and build with, and partners are supporting its development, but actual consumer-facing implementations have yet to launch. The key idea The idea behind AP2 is that, to develop a standardized way for agents to transact on behalf of users, cooperation is needed among the parties typically involved in online purchasing. As an open source project, AP2 could offer a common foundation "to securely authenticate, validate and convey an agent's authority to transact." "Having a protocol, essentially a common standard, allows for multiple agents to interact with each other," PayPal global head of AI Prakhar Mehrotra told VentureBeat in an interview. "Technically, we can do agentic payments now, but I can't guarantee the system's robustness, and I can't bring that to a regulator. A standard gives that needed layer." Its proponents said an open agentic payment protocol provides three important things: * Authorization to prove that a human user gave an AI agent permission to make a purchase * Authenticity so that merchants or sellers know an agent is truly acting based on the user's intent * Accountability so that anyone within the ecosystem knows how to deal with fraudulent or incorrect transactions Digital transactions, such as buying clothes online, operate on the same basic principles, but there is an assumption that anyone clicking the "buy" button is a human who owns the bank account or card used for the transaction. Developed as an extension of Google's earlier Agent2Agent (A2A) and Anthropic's increasingly popular Model Context Protocol (MCP) initiatives, AP2 establishes a common structure that supports diverse payment methods including credit and debit cards, stablecoins, and real-time bank transfers. By offering a payment-agnostic, interoperable framework, AP2 enables users, merchants, and financial institutions to confidently participate in agentic transactions across platforms. According to Google, the protocol is built to prevent fragmentation across ecosystems by aligning with industry standards while addressing the unique requirements of autonomous agents. While consumer use cases are a major focus, AP2 is also expected to play a role in enterprise commerce workflows. For instance, autonomous agents could manage procurement tasks, scale software licensing based on real-time usage, or interact with marketplaces like Google Cloud's for transactable services. To support implementation, Google has published full technical specifications and reference implementations in a public GitHub repository. The company encourages developers and industry stakeholders to contribute to the open standard as it continues to evolve. The state of agentic payments today When an individual or an enterprise makes a purchase online, a couple of things happen. First, the user finds and selects the product they want to buy, then adds it to their cart. After this, they choose how to pay for the transaction. Depending on the website, people will either click on "pay with card," "pay with PayPal," "pay with Buy Now, Pay Later" or their chosen automatic payment provider (i.e., Google Pay). The user then clicks a button to finish the transaction. Once that process is complete, the information is sent to the seller and the users' bank or PayPal account, and the money is exchanged. The key to all of that is a human user who manually clicks buttons. Mehrotra said the onus to prove authority, ownership, and approval is on the person, which acts as a guardrail. However, with agentic transactions, the human may have initiated the process, but a non-deterministic model made the ultimate decision to complete the transaction. Many of the companies involved in AP2 already work on projects aimed at making agentic payments efficient and secure. These projects often "make do," meaning payment providers rely on their existing technologies, which still place the onus on a human. Mastercard unveiled Agent Pay, which utilizes the company's tokenization technology to mask personally identifiable information during digital purchases. Visa partnered with Perplexity, which brings Visa's payment network to the chat platform and enables "agentic commerce." While not necessarily dealing with payments, Intuit offers several finance agents to mid-market customers so they can understand how well their products are selling. Skyfire, a San Francisco-based startup that launched in beta in 2024, also provides developers with infrastructure to equip autonomous agents with real funds and enable them to transact directly. Skyfire operates more like a payments-as-a-service platform, offering tooling to developers who want to load agents with money, enforce spending caps, and authenticate agents through an identity layer. Unlike AP2, which focuses on cryptographic proof of user intent through digital mandates, Skyfire enables users to directly fund spending accounts for AI agents using stablecoins or traditional payment methods with pre-set spending limits. Skyfire provides an out-of-the-box solution for agent payments, Google's AP2 aims to create a shared foundation that a broader ecosystem -- including players like Skyfire -- could ultimately build on. How AP2 works Mastercard chief digital officer Pablo Fourez said in an email to VentureBeat that protocols like AP2 "represent a broader industry effort to enable agent-based commerce." "Our focus is on defining the trust layer: enabling secure and responsible agentic transactions today through tokenized credentials across our global payments network, while also preparing for a future where the internet itself is rewired to support agentic commerce more natively. It's this dual focus - meeting the needs of now while building for what's next - that allows innovation to move from concept to production with confidence," he said. AP2 helps build that trust through what the consortium calls Mandates, or "tamper-proof, cryptographically-signed digital contracts that serve as verifiable proof of a user's instructions." The Mandates would be signed with verifiable credentials (VCs), which the entire ecosystem recognizes are ground truth for transactions, providing cryptographic proof that an agent's actions align with user instructions. The blog from Google said Mandates addresses the different ways people use agents to shop. * Real-time purchases with a human present, where someone asks an agent to find them something -- for example, running shoes. The request is captured by an Intent Mandate that represents the auditable context for the purchase. After the agent adds the shoes to the cart, the person approves the action with a Cart Mandate. * Delegated tasks without a human are when people tell an agent to buy, say, a concert ticket, but offer a condition, such as only buying if the ticket goes on sale. The same Mandates exist, with the Intent Mandate specifying the specific rules the user added that will trigger a purchase. This "pre-authorization" acts as the Cart Mandate. This system creates an audit trail that links a user's original request to the final payment, addressing concerns around authorization, authenticity, and liability. Paypal's Mehrotra said AP2 could bring about different ways of paying for purchases. For example, people can instruct agents to coordinate between airlines and hotels, finding the best combination of timing, price, and location to pay for an entire trip with minimal human intervention. Integrating Crypto and Web3 Payments To support the growing relevance of crypto payments, Google has introduced an extension to AP2 called A2A x402, developed in partnership with organizations like Coinbase, MetaMask, and the Ethereum Foundation. This extension makes it possible for AI agents to transact using stablecoins and other crypto assets, bringing the same structure and verifiability to blockchain-based commerce. Partners in the Web3 space welcomed the move. Coinbase described x402 as a production-ready solution that gives developers a real playground for crypto-enabled agent interactions, while MetaMask emphasized the benefit of composability and user control in decentralized environments. The adoption problem There's no doubt that payment networks, merchants, banks and others in the ecosystem will want to embrace standards like AP2. People and enterprises are understandably wary of not having much control over paying for things. It took years for people to become comfortable enough to initiate large purchases online. On a personal note, I prefer manually adding my payment information for any expensive online transactions, such as buying furniture. The US is slowly marching towards becoming cashless. Some countries are only starting to offer a Venmo-like person-to-person payment app. Standards like AP2 could go a long way to providing that layer of trust and accountability. But it is the agents themselves that need to offer that assurance to enterprises and individuals. What businesses can do with AP2 today The framework is now open for exploration and development, meaning businesses looking to integrate AP2 or prepare for agent-led commerce can take concrete steps today. First, developers and technical teams can review the public GitHub repository, where Google has published the full AP2 specification, example implementations, and supporting documentation. This resource provides a starting point for building agent payment flows that adhere to the protocol's design for secure, verifiable transactions. Second, companies building AI agents -- or integrating agents into existing platforms -- can begin designing interactions around Mandates and Verifiable Credentials, which are core to AP2's model of trusted authorization. Whether the goal is enabling real-time user-approved purchases or automating delegated tasks like travel booking or subscription renewals, these constructs offer a flexible foundation for establishing user intent and transaction traceability. Third, merchants and payment providers interested in future-proofing their systems for agentic commerce can evaluate how their existing checkout, authentication, and compliance infrastructure aligns with AP2's structure. Because the protocol is payment-agnostic and supports a range of methods (from traditional cards to stablecoins), early compatibility planning may help reduce future integration overhead. Finally, Google is encouraging community participation. Businesses are invited to contribute feedback, propose extensions, and help shape how the protocol evolves -- particularly around emerging areas like decentralized identity, risk management, and interoperability with crypto payment systems. In short, while AP2 is not yet commercially deployed, its release marks a clear opportunity for forward-looking businesses to begin building or aligning with a shared foundation for AI-driven commerce.

Google's plan to build trust in AI agents as personal shoppers

Why it matters: Industry analysts expect consumers to rapidly embrace buying stuff with the help of AI -- but only if they can count on a chatbot doing exactly what they tell it to do. Driving the news: Google's new agent payments protocol, or AP2, extends existing standards -- like the agent to agent (A2A) and model context (MCP) protocols -- to e-commerce. * Google is assuming both consumers and merchants will use "multiple agents built by multiple vendors," Rao Surapaneni, VP of Google Cloud's business applications platform, told Axios. * All these agents will need to communicate in a reliable and trusted way. Between the lines: The new payments protocol is designed to be able to "prove that a user gave an agent the specific authority to make a particular purchase," "be sure that an agent's request accurately reflects the user's true intent," and "determine accountability if fraudulent or incorrect transaction occurs," according to a new Google blog post. * The goal is to "capture the intent of the buyer" in a reliable way in new kinds of transactions where the human user isn't present but is represented by an agent, Stavan Parikh, VP of payments at Google, told Axios. Zoom out: The U.S. payments system that has evolved around e-commerce is complex and dependent on a wide variety of powerful players, including Visa, Apple Pay, Amazon, Stripe, Square (now Block), Zelle and other bank-backed systems. Zoom in: Google says AP2 already has buy-in from more than 60 partners, including Mastercard, American Express, PayPal, Coinbase, Etsy, Okta and Alibaba. Our thought bubble: An open standard backed by Google could rapidly spread, or face multiple challenges and rivals. * Single-standard systems offer convenience and interoperability, but ecosystems with multiple competing standards can be more resilient and adaptable. What they're saying: Google's success at bringing partners and competitors like Amazon and Microsoft along with its agent-to-agent protocol shows it understands an "open and collaborative environment" is the best way to build a new ecosystem, Surapaneni says. Flashback: In the early days of the web, users were reluctant to make purchases on new platforms until they could trust that the technology wouldn't run wild on them and that their credit card info and other personal data wouldn't be abused.

Google releases AI payments protocol that includes support for stablecoins, partners include Coinbase and Salesforce | Fortune Crypto

To add compatibility with stablecoins, Google worked with the crypto exchange Coinbase, which has built its own AI and crypto payments scheme. It also collaborated with other crypto companies, including the Ethereum Foundation. For other elements of the new payment protocol, Google conferred with more than 60 other organizations including Salesforce, American Express, and Etsy. "The way we built it is from the ground up to factor in both heritage and existing payment rail capabilities as well as forthcoming capabilities such as stablecoins," James Tromans, the head of Web3 at Google Cloud, told Fortune. In the buzzword-laden world of AI, "agents' are AI algorithms designed for specific use cases that can call up apps on one's device or across the internet. Some companies have created agents that specialize in writing code. Others have created AI agents that browse online marketplaces to buy, say, clothes for a human shopper. Many tech executives predict that, in the near future, AI will increasingly talk with AI, with no humans in between. If this comes to pass, AI financial advisors could talk with AI representatives at different financial institutions to find mortgages for housebuyers, or AI personal shoppers could chat with AI shopkeepers to find the perfect pair of jeans for sharp dressers. That's why Google released in April a new protocol that creates a standard means for different agents to communicate with each other. Tuesday's launch of a protocol for agent-to-agent payments builds on that framework and ensures that transactions between two AI agents are safe, secure, and what the humans on either end intended, said Tromans, the Google executive. Coinbase and Google worked together to make sure their payments schemes were interoperable, Erik Reppel, head of engineering at Coinbase's developer platform, told Fortune. "We're all working to figure out how we can make AI transmit value to each other," said Reppel.

Google Reveals AI Agent Payments Protocol Backed by Coinbase, Ethereum Foundation - Decrypt

The Ethereum Foundation says the effort aligns with its ERC-8004 standard to help AI agents securely discover, verify, and transact on Ethereum. Google took its latest step Tuesday to hasten the arrival of the AI agent economy, announcing the creation of an open-source protocol for agent payments backed by some of crypto's biggest companies. The Agent Payments protocol revealed today can be used as an extension of Google's previously released Agent2Agent protocol, which attempted to create a standard for interactions between AI agents -- automated assistants authorized to zip around the internet completing complex tasks on behalf of humans or organizations. Today's announced protocol adds the crucial ingredient of payments to that mix, with the goal of enabling AI agents to seamlessly complete financial transactions with users, merchants, and payments providers. It is backed both by crypto mainstays like Coinbase and the Ethereum Foundation, as well as traditional payments giants and vendors including PayPal, American Express, and Etsy. Other crypto firms supporting the initiative include Sui creator Mysten Labs, prominent crypto wallet MetaMask, and Eigen Labs, the firm behind Ethereum restaking protocol EigenLayer. (Disclaimer: MetaMask is a product of Consensys, one of 22 investors in an editorially independent Decrypt.) Google said Tuesday its new protocol represents an effort to create a platform- and currency-agnostic "common language for secure, compliant transactions between agents and merchants." It will support credit and debit card transactions, stablecoins, and real-time bank transfers. To accelerate the adoption of other digital assets in its AI agent framework, Google has also launched a crypto-specific extension for its Agent2Agent protocol, dubbed x402. The Ethereum Foundation was one of numerous crypto organizations that contributed to the x402 extension, which it says will complement its own efforts to make Ethereum the bedrock of the booming AI agent economy. The Ethereum Foundation is currently finalizing ERC-8004, a standard for agents to discover, verify, and transact with each other on the network, which was backed by members of Google's Agent2Agent team. "ERC-8004 will support many forms of payments, but having the x402 extension helps the developer experience," Davide Crapis, head of the Ethereum Foundation's newly announced AI team, told Decrypt. "Beyond the ERC, x402 is getting adoption, so this extension opens up a new avenue for agents built on crypto rails." Crapis added that today's announcement from Google is aligned with his team's goal of creating better infrastructure for agentic interaction on Ethereum. "It's exciting to see the idea of agents paying each other resonate with the broader AI community," Erik Reppel, head of engineering at Coinbase Developer Platform, said Tuesday in a statement. "Bringing x402 into AP2 to power stablecoin payments made sense -- it's a natural playground for agents to start transacting with each other and testing out crypto rails."

Google's Agent Payments Protocol provides a framework for AI to automate consumer purchases - SiliconANGLE

Google's Agent Payments Protocol provides a framework for AI to automate consumer purchases Google LLC is giving artificial intelligence agents even more autonomy with the launch of a new protocol that makes it possible for them to make payments on behalf of their users. The Agent Payments Protocol, known as AP2, is backed by more than 60 merchants and financial institutions, and is aimed at AI agents that can shop and make buying decisions at the request of humans. It's designed to be interoperable across AI platforms, payment systems and vendors and fully traceable so every transaction can be audited. In a blog post, Google Vice Presidents Stavan Parikh and Rao Surapaneni stressed that the company is committed to full openness and transparency, and sees the AP2 protocol evolving collaboratively over time. It hopes to attract the involvement of the "entire payments and technology community", as well as standards bodies to ensure agentic payments are regulatory compliant. Google has open-sourced the protocol and made the full specification available on GitHub in line with this stance. Parikh and Surapaneni said AP2 is designed for a future where humans routinely rely on AI agents to shop for products and services on their behalf, and envisions a world where retailers also rely on agents. AI agents have already come a long way. With the rise of browser-based agents that can surf the web and navigate through menus and other elements like people do, it's already possible for them to browse the internet and find cheap flights or hotels, and in the case of businesses, it's possible for AI agents to compare different suppliers, for example. However, today's agents still stop short of completing purchases or transactions, and instead ask users to take this final step themselves. It's this final step that Google wants to automate with AP2. In one example, Parikh and Surapaneni said this can provide a lot of convenience and allow shoppers to take advantage of the best deals available. For instance, they gave the example of a chatbot user who asks it to shop for a bicycle, which triggers a spontaneous, time-sensitive offer from a bicycle shop's own AI agent. Alternatively, someone might ask their agent to book travel and lodgings, providing it with only the dates, location and their budget. In this case, "the agent can then interact with both airline and hotel agents, as well as online travel agencies and booking platforms, and once it finds a combination that fits the budget, it can execute both cryptographically-signed bookings simultaneously." Automating such transactions are very complex, not only on the technology side, but also socially. This is why AP2 always requires AI agents to obtain two separate approvals from users before any purchase can be made. First, it needs to obtain an "intent mandate", which is where the user tells them they are looking for a new bicycle and provides an appropriate budget and specifications, enabling it to go search for a specific product and negotiate a price. Then, it also requires a "cart mandate", which is when the user gives the final approval for it to buy a specific item once it has been identified. There is a provision in the AP2 protocol for entirely automated purchases too, where the AI agent can be permitted to automatically generate the cart mandate once the specific item has been found. In such scenarios, the agent must obtain a more detailed intent mandate, which specifies price limits, the timing of the purchase and other rules. The goal is to ensure that the user retains control of their wallet and their finances and that the agents don't start making purchases willy-nilly, and also to maintain a fully auditable trail of each transaction, so it can be examined in case it's suspicious. Google has obtained a lot of industry support, with multiple payment processors getting onboard with the idea of AP2 and agentic payments. Mastercard Inc. Chief Digital Officer Pablo Fourez said the protocol can help to shape the future of agentic commerce, but stressed there's still a lot of work to be done. "These efforts include critical work with standards bodies such as the FIDO Alliance, where we are advancing verifiable credentials to capture and secure consumers' intent in this dynamic new context," he explained. "we're playing an essential role in securing the payments ecosystem - ensuring that trust and safety remain at the core of every transaction." Google is also collaborating with cryptocurrency firms such as Coinbase Global Inc., the Ethereum Foundation and the crypto wallet provider MetaMask on an extension called x402 that will enable agents to utilize digital stablecoins for purchases. "Blockchains are the natural payment layer for agents, and Ethereum will be the backbone of this," said MetaMask's AI lead Marco de Rossi. "With AP2 and x402, MetaMask will deliver maximum interoperability for developers and will enable users to pay agents with full composability and choice while retaining the security and control of true self-custody". Google isn't alone in trying to enable AI agents to make payments. Some of its rivals in the AI and payments industries are working on similar efforts. For instance, the AI search startup Perplexity Inc. offers a "Buy With Pro" service in its agentic AI browser, and Stripe Inc. has developed its own software to enable AI agents to make purchases using its platform.

Google Launches Agent Payments Protocol to Standardise AI Transactions

The protocol, developed with more than 60 payments and technology companies, extends Google's existing Agent2Agent (A2A) and Model Context Protocol (MCP) frameworks. Google on Wednesday announced the Agent Payments Protocol (AP2), an open standard designed for AI agents to conduct secure and verifiable payments. The protocol, developed with more than 60 payments and technology companies, extends Google's existing Agent2Agent (A2A) and Model Context Protocol (MCP) frameworks. Stavan Parikh, vice president and general manager of payments at Google, said the rise of autonomous agents requires a new foundation for trust. He added that AP2 establishes the foundation for authorization, authenticity, and accountability in agent-led transactions. "AP2 provides a trusted foundation to fuel a new era of AI-driven commerce. It establishes the core building blocks for secure transactions, creating clear opportunities for the industry-including networks, issuers, merchants, technology providers, and end users-to innovate on adjacent areas like seamless agent authorization and decentralized identity," Parikh said. Unlike traditional payment systems that assume a human directly initiates a purchase, AP2 addresses the challenges of proving intent and authority when an AI acts on a user's behalf. The framework uses cryptographically signed digital contracts called Mandates to serve as verifiable proof of a user's instructions. These can cover both real-time transactions, where a customer is present, and delegated tasks, such as buying concert tickets automatically under pre-approved conditions. Rao Surapaneni, vice president and general manager of business applications platform at Google Cloud, said the protocol provides secure, compliant transactions between agents and merchants while supporting multiple payment types, from cards to stablecoins. Google said AP2 will also support cryptocurrency payments through an extension called A2A x402, developed in partnership with Coinbase, Ethereum Foundation and MetaMask. This allows agents to handle stablecoin payments within the same framework. Industry players expressed support for the initiative. Luke Gebb, executive vice president of Amex Digital Labs, said the rise of AI commerce makes trust and accountability more important than ever, and AP2 is intended to protect customers. Coinbase head of engineering Erik Reppel said the inclusion of x402 showed that agent-to-agent payments aren't just an experiment anymore and are becoming part of how developers actually build. Adyen co-chief executive Ingo Uytdehaage said the protocol creates a "common rulebook" to ensure security and interoperability across the payments ecosystem. Backers include Mastercard, PayPal, Revolut, Salesforce, Worldpay, Accenture, Adobe, Deloitte and Dell, who said the framework could open up opportunities for secure agent-driven commerce ranging from consumer shopping to enterprise procurement. Google has published the technical specifications and reference implementations in a public GitHub repository and invited the wider payments and technology community to contribute to its development.

Google unveils open-source protocol for AI payments with stablecoin support

Google's AI payment protocol was developed in collaboration with Coinbase, signaling crypto's growing role in powering the AI-driven digital economy. Google has unveiled an open-source protocol that allows AI applications to send and receive payments, including transactions using stablecoins -- highlighting the growing role of dollar-pegged cryptocurrencies in the emerging AI-driven web. The initiative is being launched in partnership with Salesforce, American Express and more than 60 other companies, Fortune reported Tuesday. Stablecoin functionality was developed in collaboration with crypto exchange Coinbase, and the Ethereum Foundation was also consulted on the project. James Tromans, head of Web3 at Google Cloud, confirmed the rollout, telling Fortune that the protocol was designed to support both "existing payment rail capabilities as well as forthcoming capabilities such as stablecoins." Coinbase engineer Erik Reppel said the exchange worked with Google to make their payment systems interoperable. "We're all working to figure out how to make AI transmit value to each other," he told Fortune. The payment system builds on Google's Agent2Agent Protocol, introduced in April, which provides a framework for AI agents to exchange information and interact more efficiently. It was developed with support from more than 50 technology partners, including PayPal, Salesforce and SAP, and consulting firms such as Deloitte, McKinsey and PwC. The Tuesday announcement comes amid a growing push to connect AI agents -- autonomous software programs capable of making decisions without human input -- with decentralized finance (DeFi) protocols. Such integration could streamline trading, improve user interaction, and expand real-world payment use cases. Related: Ethereum Foundation forms AI research team to blend blockchain, AI Stablecoins may prove to be the most impactful crypto use case for AI agents, emerging just as both technologies gain broader mainstream adoption. These dollar-pegged tokens recently received a regulatory boost in the United States through the GENIUS Act. Galaxy Digital CEO Mike Novogratz recently underscored stablecoins' importance in this shift, predicting that AI agents will eventually become "the biggest user of stablecoins." The Ethereum Foundation has also highlighted stablecoins' potential in powering AI-driven applications. In August, the foundation pointed to the dormant HTTP 402 status code -- which denotes "payment required" -- and noted that, when paired with Ethereum Improvement Proposal (EIP) 3009, it could enable AI agents to execute stablecoin transfers automatically. At the time, the foundation stated that "autonomous agents are about to become Ethereum's biggest power users."

Google unveils Agent Payments Protocol to power AI commerce

This content has been selected, created and edited by the Finextra editorial team based upon its relevance and interest to our community. The tech giant has lined up more than 60 partners - including Adyen, Coinbase, Mastercard and PayPal - for the Agent Payments Protocol (AP2). Agentic commerce breaks the assumption that today's payment systems generally make that a human is directly clicking "buy" during a transaction. Google says a protocol is needed to address this, providing the basis to prove that a user has authorised an agent to make a purchase; to enable a merchant to authenticate that the agent's request reflects the user's intent; and to determine accountability if a fraud or error occurs. AP2 is designed to be an open, shared protocol that provides a common language for secure, compliant transactions between agents and merchants, helping to prevent a fragmented ecosystem. It also supports different payment types - from credit and debit cards to stablecoins and real-time bank transfers. The protocol uses mandates - tamper-proof, cryptographically-signed digital contracts that serve as verifiable proof of a user's instructions. This, says Google, addresses the two primary ways a user will shop with an agent: real-time purchases where the human is present, and delegated tasks where the person is not present. Google has also worked with Coinbase, Ethereum Foundation, MetaMask and others on an A2A x402 extension, a production-ready offering specifically for agent-based crypto payments.

Google and Coinbase Join Forces to Bring Stablecoin Payments to AI Apps

What if your next business partner wasn't human, but an AI agent capable of negotiating deals, managing resources, and executing transactions independently? In a new move, Google and Coinbase have unveiled an innovative initiative set to redefine commerce by merging artificial intelligence with blockchain technology. At the heart of this collaboration lies the x402 payment protocol, a system designed to handle seamless, low-cost microtransactions, empowering autonomous AI agents to operate in ways that were once the stuff of science fiction. This isn't just a technological leap; it's a bold step toward a potential AI-driven economy, where machines transact, collaborate, and innovate with minimal human intervention. This event recap explores the potential of this initiative and the infrastructure powering it, from agentic marketplaces to the x402 protocol. You'll discover how Google's AI expertise and Coinbase's blockchain leadership are allowing autonomous systems to thrive in secure, scalable ecosystems. But it's not all smooth sailing; this new frontier raises questions about monopolization, security, and equitable access. As you read, consider how this shift could reshape industries, challenge traditional business models, and spark debates about the role of humans in an increasingly automated world. By combining Google's expertise in AI and Coinbase's leadership in blockchain technology, this partnership seeks to unlock new possibilities in automation, efficiency, and innovation. The initiative not only highlights the potential of autonomous AI systems but also underscores the critical role of blockchain in creating secure and scalable ecosystems for the future of commerce The AI agent economy is a concept that envisions autonomous AI systems operating independently to perform tasks, pay for services, and interact with other agents. These AI agents are designed to function without human intervention, executing transactions and coordinating activities at speeds and scales that far exceed human capabilities. While promising, much of this remains aspirational and in the testing stage, supported by frameworks like Google's AP2, which is currently open to developers and partners but not yet deployed for consumers. To support this vision, Google and Coinbase have introduced early controlled environments where AI-driven transactions can occur in isolation from traditional financial systems. This approach ensures that potential risks are minimized while allowing AI agents to operate within a secure and structured framework. These controlled spaces serve as a testing ground for refining the capabilities of AI agents, allowing them to interact seamlessly and efficiently in a safe setting. The x402 payment protocol is the cornerstone of this emerging AI commerce ecosystem. Developed by Coinbase, x402 uses blockchain technology to enable fast, low-cost, and highly divisible microtransactions using stablecoins. Unlike traditional payment methods, such as credit cards or bank transfers, x402 is specifically optimized for the high-frequency, low-value transactions that are characteristic of AI agent interactions. By using blockchain, the protocol ensures transparency, security, and reliability in every transaction. This makes it particularly well-suited for applications such as: The x402 protocol's ability to handle microtransactions efficiently is a critical enabler for the potential AI agent economy, allowing autonomous systems to interact and transact without the friction of traditional financial systems. Advance your skills in AI Agents by reading more of our detailed content. Agentic marketplaces are virtual platforms where AI agents could one day autonomously discover, negotiate, and interact with services. These marketplaces, such as Google's A2A (Agent-to-Agent) protocol and the x402 Bazaar, provide the infrastructure needed for AI agents to collaborate and achieve objectives with limited human involvement. For example, an AI agent might in the future independently book travel arrangements, conduct background checks, or purchase goods and services. These platforms are intended to streamline interactions between AI systems, allowing them to operate efficiently in a shared ecosystem. Key features under development include: These marketplaces represent a significant advancement in AI-driven commerce, but much of their capability remains potential rather than widely available reality today. The potential applications of the x402 protocol and AI agent economy span a wide range of industries and use cases. Some of the most promising examples include: These use cases demonstrate the versatility and potential of AI-driven commerce, offering new opportunities for efficiency and innovation across various sectors. While the AI agent economy holds immense promise, it also presents several challenges that must be addressed to ensure its success. Key concerns include: These challenges also present opportunities for innovation and collaboration. By addressing these issues proactively, developers and businesses can create robust, inclusive, and secure AI-driven services. Early adopters of this technology are likely to gain significant competitive advantages, much like the pioneers of previous technological revolutions. The collaboration between Google and Coinbase marks the beginning of a new era in commerce and automation. As the x402 protocol and AI agent economy continue to evolve, the potential for fully autonomous businesses and AI-driven ecosystems becomes increasingly tangible. However, realizing this vision will require the development of clear regulations, robust infrastructure, and collaborative efforts among policymakers, developers, and businesses. It is important to note that AP2 is not yet fully deployed for consumer use. As reported by VentureBeat, the protocol is currently being opened to developers and partners for experimentation, with broader adoption still to come. The societal and economic shifts anticipated from these advancements will demand careful navigation to maximize benefits while mitigating risks. By fostering innovation and addressing challenges, the AI agent economy has the potential to reshape industries, redefine business models, and unlock new opportunities for progress.

Google Unveils a Payment Protocol for AI-Driven Commerce | PYMNTS.com

AI agents are shifting from chat to checkout, challenging a core assumption of eCommerce: that a human presses "buy." AP2 proposes a consistent way to prove that an agent's request reflects a user's instruction and to identify who approved the transaction. Google said more than 60 firms are supporting the launch, including American Express, Mastercard, PayPal, Etsy, Revolut and Coinbase. The mechanism rests on tamper-evident mandates, cryptographically signed records that travel with each transaction. In a human-present flow, the agent captures an intent mandate when the user requests an item, followed by a cart mandate that approves the final basket. A third payment mandate signals to networks that an AI agent executed the order. Together, these credentials create a non-repudiable trail that issuers and merchants can use in fraud investigations and dispute resolution (AP2 GitHub Spec). Google illustrated scenarios that highlight AP2's role. A shopper might direct an agent to monitor a jacket in a specific color and buy if the price falls below a threshold. A traveler can set a budget and let the agent coordinate flights and hotels, then approve both bookings at once. For unattended purchases, such as buying concert tickets at release, a detailed intent mandate sets rules upfront, allowing the agent to generate a cart mandate later without a human present. While the initial version focuses on card rails, AP2 was built to be payment-agnostic. Google released an extension linking the protocol to the x402 crypto standard, developed with Coinbase, MetaMask and the Ethereum Foundation, so agents can transact directly from crypto wallets. Coinbase published an explainer on the integration, positioning it as a way to unify AI commerce with decentralized payments. That roadmap signals where Google expects AI-driven payments to go. Future iterations of AP2 are slated to include real-time bank transfers and stablecoin support, bridging the gap between traditional card networks and instant push-payment systems. Google is not alone in preparing for AI commerce. Mastercard is rolling out its Agent Pay service globally, part of a toolkit that helps merchants and developers create agent-ready checkout flows. Visa, meanwhile, has opened its MCP server and released a no-code Agent Acceptance Toolkit so developers can plug agents into Visa APIs and test real transactions. Platforms are experimenting as well. In May, PayPal partnered with Perplexity to power its "Buy With Pro" service, embedding agent-led purchases inside its browser. The move signaled that agentic checkout is no longer theoretical; it is routing real orders. These examples show why a standard like AP2 matters: it gives merchants and issuers a uniform way to handle agent-driven requests rather than negotiating bespoke integrations. For risk teams, AP2's biggest contribution is accountability. Each mandate documents what the user allowed, what the merchant promised, and what the network processed. That evidence can cut down on false disputes and help issuers make better approve-or-decline decisions. Agentic commerce will still need strong identity checks, but mandates could reduce fraud without adding friction. Adoption will hinge on two tests. First, developer uptake. The AP2 GitHub repository contains full specifications, sample flows and reference code, lowering the barrier for pilots. Second, merchant economics. If AP2 reduces fraud losses and false declines, or lifts conversion by enabling richer agent-to-agent offers, merchants will adopt. If it adds steps without visible upside, they may wait. The final question is regulatory and network acceptance. If mandates are recognized as sufficient proof of authority, AI-driven transactions can scale through existing payment pipes. If not, each platform may need custom agreements. For now, AP2 provides a foundation: it shifts proof of purchase from inference to verifiable, signed intent, giving the ecosystem a common starting point to build on.

Google's New Agent Payments Protocol (AP2) : The Future of AI Shopping Without Lifting a Finger

What if your next online purchase didn't require you to click a single button, or even be present? Imagine a world where autonomous agents handle everything from snagging limited-edition sneakers the moment they drop to making sure your pantry is always stocked, all while maintaining your privacy and security. With the unveiling of Google's Agent Payments Protocol (A2P), this vision is no longer science fiction. A2P is more than just a technological leap; it's a reimagining of how we interact with money in an era increasingly dominated by automation. By empowering digital agents to execute transactions on your behalf, Google is setting the stage for a future where convenience meets innovative innovation. But how does A2P work, and what does it mean for you? Sam Witteveen explains the core of A2P, from its open source framework to its emphasis on trust and security. You'll discover how it's designed to handle everything from automated purchases to price monitoring, all while keeping you in control. Whether you're a tech enthusiast, a merchant, or someone curious about the future of financial automation, A2P offers a glimpse into a world where autonomous transactions aren't just possible, they're practical. As we unpack its potential, consider this: could this protocol redefine the way we think about money and autonomy? The Agent Payments Protocol (A2P) is a framework that enables autonomous agents to execute financial transactions on behalf of users. These agents are capable of performing tasks such as automated purchases, price monitoring, and managing repetitive transactions, significantly reducing the need for constant user input. Built upon Google's existing A2A and MCP frameworks, A2P ensures compatibility with established systems while maintaining an open and non-proprietary design. This open approach emphasizes user privacy and provides clear guidelines for liability and security, making it a robust foundation for agent-based financial systems. By allowing agents to act within predefined parameters, A2P ensures that users retain control over their financial interactions while benefiting from enhanced convenience and efficiency. A2P is designed to address practical scenarios where autonomous agents can enhance efficiency and simplify financial processes. Its applications span a variety of use cases, including: These use cases demonstrate how A2P can streamline everyday tasks, allowing users to focus on higher-priority activities while their agents handle routine financial responsibilities. Here are more guides from our previous articles and guides related to AI Agents that you may find helpful. Security and trust are at the core of A2P's design. The protocol employs cryptographic signing to validate the authenticity of transactions, making sure that agents operate strictly within their authorized scope. Each transaction includes a cryptographically signed intent mandate, which serves as a digital confirmation of the agent's authority to act on behalf of the user. A2P also distinguishes between "human present" and "human not present" transactions, addressing varying levels of user involvement and the associated risks. This differentiation allows the protocol to adapt to diverse scenarios, from fully automated purchases to transactions requiring user approval. Furthermore, A2P clearly defines the roles and responsibilities of agents, merchants, and payment processors, fostering accountability and trust among all parties involved. By incorporating these security measures, A2P not only protects users from unauthorized actions but also builds confidence in the broader adoption of agent-based financial systems. For merchants, A2P offers a streamlined approach to handling transactions, reducing friction in the purchasing process. However, the protocol also presents challenges, such as the potential reduction in opportunities for upselling or engaging customers through loyalty programs. To address these concerns, A2P integrates fraud prevention measures and identity verification systems, making sure that agent operations remain secure and transparent. The protocol also supports the broader payment ecosystem by providing tools for seamless integration. Developers can access open source resources, including comprehensive documentation and sample code, on platforms like GitHub. This flexibility enables businesses to customize A2P to meet their specific needs, fostering innovation across a wide range of industries. By simplifying the integration process and addressing security concerns, A2P encourages merchants and developers to explore new possibilities for using autonomous agents in their operations. The introduction of A2P marks a significant step toward the evolution of advanced agent ecosystems. One potential development is the creation of agent app stores, where users could browse and download specialized agents tailored to specific tasks. These platforms could expand the capabilities of autonomous agents, driving innovation in e-commerce, logistics, and beyond. A2P's emphasis on secure and efficient financial interactions also lays the groundwork for broader adoption of autonomous agents across various industries. As these systems continue to evolve, they may redefine the boundaries between human and machine transactions, raising important questions about trust, liability, and the ethical implications of automation. The protocol's open source nature further accelerates its potential for growth. By inviting collaboration and feedback from the developer community, A2P ensures that it remains adaptable and relevant in an ever-changing technological landscape. Google's decision to make A2P an open source initiative underscores its commitment to fostering collaboration and innovation. Developers can access detailed documentation, sample implementations, and other resources on GitHub, allowing them to adapt the protocol to a wide range of applications. This open approach encourages experimentation and invites feedback from the global developer community. By using the collective expertise of developers, A2P can evolve to address emerging challenges and opportunities in the realm of autonomous transactions. This collaborative model ensures that the protocol remains a dynamic and versatile tool for advancing agent-based systems.

Google Debuts Agent Payments Protocol to Bolster AI Commerce | PYMNTS.com

By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions. AP2 is designed to "securely initiate and transact agent-led payments across platforms," according to a Tuesday (Sept. 16) company blog post. Google is collaborating on agentic payments with more than 60 companies, some of which include Adyen, American Express, Mastercard, PayPal, Coinbase and Revolut, per the post. "AI agents are capable of transacting on behalf of users, which creates a need to establish a common foundation to securely authenticate, validate and convey an agent's authority to transact," the post said. "While today's payment systems generally assume a human is directly clicking 'buy' on a trusted surface, the rise of autonomous agents and their ability to initiate a payment breaks this fundamental assumption and raises critical questions that AP2 helps to address." The questions are authorization, or proving that a user gave an agent authority to make a specific purchase; authenticity, or allowing merchants to be sure an agent's request reflects the user's intent; and accountability in cases of fraud or incorrect transactions, per the post. The protocol can be used as an extension of the Agent2Agent (A2A) protocol and Model Context Protocol (MCP). In conjunction with industry rules and standards, it offers a payment-agnostic framework for users, merchants and payments providers to transact across all types of payment methods, the post said. PYMNTS Intelligence's August edition of The Prompt Economy Tracker® Series explored the rise of MCP, an open standard that was introduced by Anthropic in late 2024 and has since been adopted by OpenAI, Microsoft and Visa. "MCP is the digital equivalent of USB-C for agents," the report said. "It defines how agents plug into data, invoke APIs, talk to other agents, and complete tasks securely and efficiently. This is the infrastructure that transforms agents from smart tools into autonomous actors inside the commerce ecosystem." Meanwhile, a July PYMNTS Intelligence report, "Payments Execs Say AI Agents Give Payments an Autonomous Overhaul," revealed that agentic AI could demand new infrastructure, trust frameworks and corporate oversight. AI agents require real-time, scalable, secure infrastructure, the report said. Legacy systems can't deal with thousands of concurrent autonomous agents acting on APIs, analyzing data and triggering actions across systems.

Google partners with Coinbase for stablecoin payments in AI apps

Google has unveiled a new, open-source Agent Payments Protocol (AP2) that allows AI applications to send and receive payments -- including support for stablecoins, digital currencies pegged to fiat currencies like the US dollar. Google worked closely with Coinbase, as well as the Ethereum Foundation, and more than 60 other companies such as Salesforce, American Express, Etsy, and PayPal, to develop this protocol that integrates crypto and traditional financial rails. The protocol is an extension of Google's previously launched Agent2Agent framework, which enables standard communication between autonomous AI agents; now, it expands this to secure financial transactions. AP2 supports various payment methods: from credit/debit cards and bank transfers to stablecoins. By integrating stablecoins within AI apps, AI agents can independently execute payments -- think, for example, of an AI that automatically books concert tickets as soon as sales open.