Author: @BlazingKevin_, the Researcher at Movemaker

Google's recent AP2 agreement sets down the bottom rules for payments and transactions for the upcoming Agent economy. Its core task is to complete the last piece of the “AI analysis” (based on MPC) and “Agent execution” (based on A2A), “value settlement”, which attempts to standardize the payment process standards of AIAgent, and to model the interactive rules of Agent, and ultimately promote AI Agent, which can carry out business activities on behalf of the user。

The primary application scene of AP2 is aimed at the core of traditional Internet sites such as electricians and subscription services. This means that it was originally designed to address the issue of how to safely delegate payment authority from human beings to AI under the existing commercial framework. This will have a direct impact on the direction of electrician operations, and users will be liberated from clicking on “purchase” and “subscription” operations to “intentional commission” and “rule-setting” of AI Agent。

It's this logic that we need to assess more carefully the impact of AP2 on the encryption industry. It was not created for the original Web3 scene, but rather incorporated encrypted payments (especially stabilization currency) as an optional, integrated payment option into its macro-frame. As a result, AP2 provides an interface for encrypted assets to enter mainstream AI applications rather than trigger an AI-plus outbreak。

UNDERSTANDING AP2 PRE-AGREEMENTS: MCP AND A2A

AP2 EXISTS ON THE PATH OF THE MCP AND A2A PRE-AGREEMENTS. UNDERSTANDING THE NATURE AND LIMITATIONS OF THESE TWO AGREEMENTS IS THE KEY TO UNDERSTANDING THE REAL INTENTIONS OF AP2. MCP SOLVES THE CONNECTION BETWEEN AI AND THE TOOL, WHILE A2A TRIES TO SOLVE THE COLLABORATION BETWEEN AI AND AI。

---

WHAT'S MCP? GETTING AI INTO REAL LIFE

The essence of a large language model (e.g. Gemini, GPT) is a “super brain” isolated from the real world. It has vast static knowledge, but cannot access real-time information or directly operate any external software. Users issued instructions such as “Call me a ticket to Osaka tomorrow”, which cannot be completed because it does not have the ability to interact with the booking website。

MCP (Model subject Protocol)THE AGREEMENT WAS PROPOSED TO ADDRESS THE ISSUE OF HOW AI COMMUNICATES WITH THE OUTSIDE ENVIRONMENT. MCP HAS DEFINED A STANDARDIZED COMMUNICATION FORMAT THAT ALLOWS AI TO SEND REQUESTS TO EXTERNAL TOOLS (E.G. DATABASES, API) AND RECEIVE RETURNS. IT CAN BE UNDERSTOOD AS THE INSTALLATION OF A TELEPHONE IN THIS "SECRET ROOM BRAIN"。

The MCP (Model Context Protocol) agreement was designed to break the "standard walkie-talkie system" that existed in the room。

It is driven by industry giants such as Anthropic and aims to address how AI interacts with external data and tools (such as databases, software, block chains). Through this walkie-talkie set, the AI model enables requests to be sent to the outside world in a standard language, while the MCP system is responsible for the accurate transmission of these requests to the corresponding external tools (e.g. weather API, booking site API) and, upon completion of the tool, returns the results to AI in a standard format。

I don't knowANALYSIS OF THE CURRENT MCP REALIZATION SHOWS THAT IT'S MORE LIKE AN INSIDE PHONE THAT CAN ONLY DIAL A FEW PREDEFINED NUMBERS。

Currently, AI ' s ability to access external tools is highly dependent on vertical integration by the platform. For example, Gemini, when searching is required, calls Google Seach by default and only; its code execution also operates in the platform ' s own sandbox environment. This model has a direct impact on users and developers:

• For users: AI caps locked in platform. Users are unable to choose third-party tools that they consider to be of higher quality (e.g. use Perplexity for AI searches, or call for specific financial data API). The usefulness of AI is thus greatly limited, and user experiences are tied to the ecology of the platform。

• For the developers: This creates a de facto market barrier. Even if the developers create an AI application for a remote platform-borne tool in an area, it cannot be fairly detected and called by mainstream AI Agent. Innovations have been suppressed and market competition limited to a closed platform-led system of cooperation。

A2A: CONCEPTS FROM “INLINE TELEPHONE” TO “PUBLIC YELLOW PAGES”

If MCP is the "hands and feet" that give AI the operational tools, then the A2A (Agent-to-Agent) protocol is for independent AIAgents to learn how to communicate and collaborate. It's based on the MCP. It defines a common set of rulesIndependent AI Agent, developed by different companies and developers, can discover each other ' s abilities, understand each other ' s capabilities and work together to accomplish complex tasks that individual Agent cannot do。

The significance of A2A is to upgrade MCP's task-handling model from “single combat” to “project subcontracting”. For example, when the user raises the complex need to “preparate a beach party in Hawaii next month”, under a closed MCP-only system, the individual Agent can only timidly call on the limited tools allowed by the platform。

But in the concept of A2A, the individual Agent can play the role of “contractor”. It disassemblys the mission and then publishes the needs in the A2A network, looking for professional “travels”, “eats” and “mobilizations” to work together. This paradigm shift will have far-reaching implications:

• For users: AI Agent 's capacity will result in an exponential leap. Users can safely assign extremely complex, cross-cutting tasks, as AI is no longer a simple operator of tools but a coordinator of resources。

• For the industryA2A aims to create an open, interoperable market for Agent services. It is committed to addressing three core issues:How(Creating mechanisms similar to the Agent app store)How(Definition of standardized descriptive language for Agent capabilities) andHow to collaborate(Procedure to regularize task assignments and synchronize progress). As soon as the system matures, any eligible Agent developer can register his or her services and be fairly selected by the market。

AAP2 Core Mechanism: Building Confidence and Authorization Framework for Agent Transactions

The MCP and A2A agreement resolved the issue of Agent's "what to do" and "how to work together", but it immediately posed a more difficult business problem: who is responsible for the legality and consequences of these acts when Agent begins to operate autonomously in relation to funds

The traditional online payment system, whose wind logic is based on the basic premise that there is an active, real-time operation by a human user in front of the screen. Any non-human-initiated, automated payment request will be marked as high-risk behaviour by the existing financial system. This is the most fundamental obstacle to the economic formation of Agent. In particular, an AI-initiated transaction request raises three questions for merchants and payment networks that cannot be answered:

1. Certification of authority: HOW DO YOU CONFIRM THAT AI ACTUALLY AUTHORIZED THIS TRANSACTION
2. I want to be sure: DOES AI ' S SPECIFIC REQUEST (E.G., QUANTITY OF PURCHASES, PRICE) ACCURATELY REFLECT THE TRUE INTENT OF THE USER, ESPECIALLY IN THE ABSENCE OF REAL-TIME MONITORING BY THE USER
3. Attributability: Who should bear the loss if the transaction goes wrong

The Google AP2 protocol was designed from the very point of departure to respond systematically to these three questions. AP2 To create an open, standardized set between users, AI Agent and businessesAgreement on Authorization and Trust CommunicationI don't know。

THE CORE TECHNOLOGY OF AP2 IS A DUAL MANDATE MECHANISM:

1. Real-time authorization (Cart Mandate)

THIS MECHANISM IS NOT NEW TO CREDIT CARD USERS, AND WHEN YOU USE CREDIT CARD PAYMENTS ON YOUR WEBSITE, YOUR CHANCE TO JUMP OUT OF THE AUTHORIZATION WINDOW. THAT'S WHAT AP2 DOES.

Upon completion of previous work on commodity studies, price comparisons, etc., Agent produces a “shopping car” containing all the details of the transaction and presents it to the user. The execution of the transaction is subject to final review and signature confirmation by the user。

Impact analysis: Real-time authorization is the initial form of Agent's payment, which lowers the user's threshold, but does not fully exercise Agent's autonomy. Its main value lies in optimizing the pre-transaction decision-making process, but human intervention is needed in the end。

2. Delegation of authority (Intent Mandate)

This is the transformative part of AP2. Users can pre-set a complex set of conditional commissioning instructions. For example: “Only when the price of Tesla Model Y is reduced by more than $20,000, the purchase process is implemented automatically by using A account funds or B account if A accounts are insufficient.”

IN ORDER TO BE CREDIBLE AND SECURE, THE AAP2 REALIZATION LOGIC IS BASED ONVerified certificates, VCs:

• THE USER'S COMPLEX COMMISSIONING CONDITIONS ARE TRANSLATED INTO A DIGITIZED, TAMPER-PROOF “TRADING CONTRACT” (I.E. VC) SIGNED BY THE USER。
• AI Agent, in each of the key steps of its mandate, needs to present this “contract” to justify its conduct。
• THE CORRESPONDING VC RECORDS ARE GENERATED AT ALL STAGES OF IMPLEMENTATION, ENSURING THE AUDITABILITY OF THE ENTIRE PROCESS。

For business and clearing networks (e.g. Visa): They received a “transaction contract” with the user's encrypted signature, specifying the scope and terms of the authorization. By validating the contract, they can gain sufficient confidence to confirm the legitimacy of the transaction, thereby significantly reducing the risk of fraud and the cost of compliance。

For usersthis makes it possible to “7x24-hour autonomous economic agents” in the true sense, freeing users from high frequency, repetitive decision-making。

In short, the real innovation of AP2 is not to transform networks like Visa or stabilize currency, but above themAdded a semantic level of trust on “who spends money, what spends money, can be traced back beyond what is authorized”I don't know。

It was committed to resolving the issue of confirmation of AI ' s intention to pay in a uniform manner in different liquidation systems, such as French currency and encryption, in order to dispel fundamental concerns of all participants about the uncontrollable and unverifiable behaviour of Agent。

x402 Extension: A service call process for encoded payments embedded in Agent

AP2's core technology is designed to make AI Agent compatible with existing, human-centred financial empowerment systems, while the "A2A x 402" extension is designed for future primary, chain A2A payments. X402 Outreach promoted by Google and Coinbase and the Etherwood Foundation。

x402 Technical selection, derived from a long-term inactive Internet standard, HTTP 402 status code, defined as “Payment Required” (payment required). In the past, this status code has almost never been applied in practice due to a lack of standardized machine means of payment。

However, the value of this protocol is highlighted when the API's main caller is transformed from human developers to high frequency, automated AIAgent。

The traditional subscription or advance fee model is too heavy and inefficient for Agent to be on demand, dynamic, cross-platform to call for big services. The core idea of x402 is precisely to solve this problem, which combines the API call with the payment。

Work processes are designed to be very simple and automated:

1. AI Agent calls for a fee-paying service API。
2. The server returned directly to one402 Payment RequiredResponse, which contains the information required to complete the payment (e.g., receipt address, amount, currency type)。
3. Agent's built-in wallet or payment module resolves this “payment bill” and automatically completes payment on the chain (e.g. USDC)。
4. Agent carries a certificate of payment on the chain, restarts the API request。
5. The server validates the payment and returns the data or service results required for Agent immediately。

This process, which appears to be simple, can have an important impact on the business model of AIAgent services by taking advantage of the real-time, high programmable characteristics of stable currency settlement。

Impact on AI Agent:

• Achieving real “pay-as-you-go”: Agent does not have to go through the traditional manual process of "Registration Account - > Bonding Credit Cards - > Select Subscription Package - > Waiting for Service to Open." It can dynamically find in the network any service that follows the x 402 standard and completes payment and use in milliseconds, which provides the basis for Agent ' s autonomy and exploration capabilities。

• Extreme cost-effectivenessAgent's payments are far more frequent than humans, and parallel mission processing capabilities enable it to make micro-payments and current payments of very fine particles. x402 enables Agent to accurately pay for the amount of returned data (Token) by the number of API calls and the calculated length of time, to perfect the link between resource consumption and cost and to avoid the waste of resources under the traditional subscription model。

IMPACT ON AI SERVICE PROVIDERS:

• The payment system's protocol sunk: x402 directs the ability to “access or quote, pay or service” to the communication protocol level. Developers no longer need to build their own complex billing, account and subscription management systems, and they can focus on the development of core services by priced and marketed any API, one data fraction, or even one page component。

• Optimizing cash flows and global settlements: WITH A STABLE CURRENCY, SERVICE PROVIDERS ARE ABLE TO OBTAIN IMMEDIATE, LOW-COST GLOBAL SETTLEMENTS THAT ARE COMPLETELY REMOVED FROM THE TRADITIONAL CROSS-BORDER PAYMENT OF LONG SETTLEMENT CYCLES AND HIGH HANDLING FEES. THIS HAS PROVIDED UNPRECEDENTED FACILITIES FOR UHF, VERY SMALL TRANSACTION RECONCILIATIONS AND HAS SIGNIFICANTLY REDUCED THE THRESHOLD FOR INDIVIDUAL DEVELOPERS AND SMALL TEAMS TO COMMERCIALIZE AI SERVICES。

Summarizing and looking forward: parallel to the twin tracks, ecological emergence

THE VALUE OF THE AP2 PROTOCOL IS NOT THAT IT APPEARS TO BE A SLIGHTLY LESS TECHNICAL DETAIL, BUT RATHER THAT IT PROVIDES THE FINAL PUZZLE OF THE COMMERCIAL CLOSED CIRCLE FOR THE MCP (CONNECTION)+A2A (COLLABORATION) TECHNOLOGY EVOLUTION ROUTE --Value settlementI don't know。

By giving Agent a reliable trading capacity, it has laid the first standardized track for the entire industry to move from an "AI tool" era to a truly "Agent economy"。

This pattern of Google clearly demonstrates its intention: on the one hand, to accelerate the development of industry standards in mainstream settings such as electricians, subscriptions, etc., through the compatibility of traditional commerce,ENABLING THE RAPID “ADAPTATION” OF ELECTRICAL PLATFORMS TO AIOn the other hand, criteria for payment by Agent-to-Agent on the chain have also been established through an encapsulation programme。

Opportunities and challenges: Encrypted Junction of Argentina

The completion of AP2 capacity to pay will undoubtedly unlock a large number of applications, such as 7x24 hours of self-management, enterprise automated procurement and renewal costs, which will facilitate the penetration of the Agent-to-Agent economic form。

HOWEVER, AP2 IS AN ENABLING FRAMEWORK DRIVEN FROM THE TOP DOWN BY THE TECHNOLOGY GIANT, WITH THE OVERARCHING OBJECTIVE OF ADDRESSING ITS CORE COMMERCIAL ECOLOGY. ENCRYPTED PAYMENTS PLAY A ROLE THAT IS “COMPATIBLE” AND “SUPPORT”。

This is a positive stimulus for the encrypted original Agent protocol. At present, the Agent products that really revolve around the core values of Web3 such as decentrization, resistance, privacy protection, etc., have not yet achieved critical technological breakthroughs, much less a clear PMF。

AND WHO COULD BUILD A PRODUCT WITH ORIGINAL ENCRYPTION AND INNOVATIVE DESIGN AT A TIME WHEN INDUSTRY STANDARDS ARE BECOMING CLEAR IS THE WATERSHED OF THE NEXT COMPETITION FOR ENCRYPTED AI PROTOCOLS.

ALTHOUGH THE WAY FORWARD IS NOT CLEAR, SOME OF THE AGREEMENTS IN THE ENCRYPTION INDUSTRY HAVE BEGUN TO BE ACTIVELY STRUCTURED AROUND AP2 AND THE POSSIBILITIES IT OFFERS. FROM OUR FIRST PARTNERS, WE CAN SEE AN INITIAL ECOSYSTEM EMERGING:

Standards and infrastructure: Coinbase, as the original proposer of the x402 payment standard, is actively promoting its landing; and the Etherwood Foundation is working to develop bottom-up agreements such as ERC-804 (trust agency criteria)。

Entry and Wallet: MetaMask works to build self-custody AI proxy wallets and to simplify the user entry process in an attempt to become a secure entry point for the Agent economy。

The public chain and interoperability: Mesh focuses on optimizing payment routes to ensure the success and efficiency of Agent payments。

Payments and applications: Crossmint, BVNK, etc., provide multiple channels of encryption and French capacity to pay for Agent; and QuestFlow, for example, already has an internal x 402 micropayment system that allows Agent to be paid according to the outcome of the mission。

Ultimately, the release of AP2 marked the beginning of the process of standardization of the Agent economy. It has created an unprecedented imagination for the industry, but at the same time has further concentrated power to control these new economic rules。

For the encryption industry, it is both an excellent opportunity for mainstreaming, accelerated adoption and a critical test of future Agent economic dominance. How to use this infrastructure while moving out of a differentiated path with Web3 core values will be a noteworthy track in the coming year。

About Movemaker

Movemaker is the first official community organization to be launched jointly by Ankaa and BlockBooster, under the authority of the Aptos Foundation, and focuses on promoting the construction and development of the Aptos Chinese-speaking community. As the official representative of Aptos in the Chinese-speaking Community, Movemaker aims to build a diverse, open and prosperous Aptos ecosystem by connecting developers, users, capital and numerous ecological partners。

Declaration of exemption:

This post/blog is for reference only and represents the author's personal point of view and does not represent the Movemaker's position. It is not intended to provide: (i) investment proposals or investment recommendations; (ii) offers or solicitations for the purchase, sale or holding of digital assets; or (ii) financial, accounting, legal or tax recommendations. Holding digital assets, including stable currencies and NFTs, is extremely risky, with high price volatility and may even become worthless. Carefully consider whether a transaction or holding of a digital asset is appropriate for you, depending on your financial situation. Ask your legal, tax or investment adviser if you have any specific problems. The information provided in this paper (including market data and statistical information, if any) is of general interest only. Due diligence has been exercised in the preparation of these data and charts, but no responsibility exists for any factual errors or omissions expressed therein。