The World Cup offers a concrete setting for observing how Web3 wallets are changing. Topics that fans discuss every day, such as whether a team can qualify or how championship odds shift, become priced and tradable behavior in prediction markets such as Polymarket. When mainstream Web3 wallets integrate World Cup prediction activities, the experience can be seen as a lightweight entry point for users to begin interacting on-chain over a longer timeline. The source also connects this setting with the extended reading titled “世界杯狂欢，预测市场上桌：Polymarket 们如何撕开大众化口子？”
imToken brings agents into World Cup prediction flows
In this World Cup-related exploration, imToken has started putting AI agents into real usage scenarios. Its web-based agent and Discord agent are designed around specific prediction needs, helping users complete betting transactions in a more natural way. Users no longer have to operate only inside the wallet app. They can participate in prediction markets from Discord or a web page, while the agent guides them back to on-chain interaction with less friction. This is an early form of the Agentic Wallet: a future Web3 wallet may not be limited to a single wallet application, but may instead appear as an AI wallet format distributed across multiple user environments.
For the past decade, the core questions that wallets solved were clear: where assets are stored, who controls the private key, and who signs the transaction. When users open imToken, they usually want to check balances, transfer assets, interact with protocols, or manage multi-chain holdings. In that stage, the wallet functions mainly as an asset entrance and a signing entrance. The user already knows what they intend to do, and the wallet completes the final step.
The Agentic era changes this flow because users may not know what button to click at the beginning. In a World Cup scenario, an ordinary user may not first think: open Polymarket, find a specific market, evaluate the odds, and execute a trade. The more natural questions are: What is interesting about tonight’s match? I favor Portugal to qualify; is there a related market? Are the odds already too low? If I only want to take part with a small amount, how should the process work? In the past, these questions were scattered across WeChat groups, social platforms, or search engines. Each user had to piece the information together and then execute step by step. After an agent enters the process, the interaction changes: the user expresses a general intent, the agent breaks down the route, and the wallet turns that route into a sequence of on-chain actions.
Wallet boundaries extend beyond the app
This is not simply a matter of adding a chat box to a wallet. The real change is that the wallet moves from being a function menu to being an intent interpreter. Earlier wallets mostly waited for the user to decide whether to transfer, swap, stake, or connect to a DApp. A newer wallet experience lets the user describe the desired result in natural language. A mass event such as the World Cup is suitable for this because it already has context, gives users a reason to express opinions, and requires judgment. The agent does not need to manage complex portfolios from the beginning, especially because complex asset management carries higher risk. It can first help users find an interaction path inside a specific scenario, and then return final control to the wallet and the user.
The web and Discord agents mentioned in the source are representative examples. They bring wallet capabilities into lighter entry points. A user does not necessarily open the app first, and does not necessarily enter a traditional DApp browser first. Instead, the user can start from an event page or a World Cup context and use the agent to find the correct interaction path. In this sense, the boundary of the wallet is spilling outward. Earlier wallet entrances were relatively clear: open the app, enter the asset page, click a function, and connect to a DApp. Future wallet entrances may be scattered across web pages, Discord, Telegram, AI chat boxes, campaign pages, developer tools, or even lightweight wallet interfaces generated by users themselves.
From this angle, the World Cup prediction activity is not only about prediction markets. Its larger role is that it allows the wallet to stand more naturally in front of the user’s intent. If the discussion remains only inside the crypto industry, Agentic Wallets can easily be treated as another narrative in which AI helps users read markets, find opportunities, or place trades. That sounds close to the previous wave of AI agent enthusiasm. But the launch of Mastercard’s Agent Pay for Machines on June 10 moves the topic beyond Web3 and into a broader payment context.
Mastercard and WeChat Pay show how payment boundaries are being designed
Mastercard’s definition of Agent Pay is specific: trusted AI agents should be able to participate in payments under user authorization. The design covers how an agent is identified, how it is authorized, how it is verified inside a payment network, and how merchants, issuers, and users know that a transaction was assisted by an agent. These are highly similar to the questions Web3 wallets face. When AI only helps a user write copy, the cost of an error is usually manageable. Once AI starts participating in asset interaction, the questions change: Does it have permission? Has it understood the user’s intent correctly? Is the service it calls trustworthy? Has the transaction it initiates exceeded the boundary? If the result differs from what the user expected, how is the outcome handled?
Mastercard’s answer is to redesign the identity, tokens, authorization, risk control, and dispute handling of trusted agents inside the payment network. The fact that a traditional financial giant is designing payment infrastructure for Agentic Commerce shows that the issue has entered a more practical commercial setting. A similar direction is also appearing among domestic payment companies. WeChat Pay is testing AI payment functions together with Tencent’s agent product WorkBuddy, including an “AI 专属卡” inside WeChat Wallet. Based on the disclosed information, the core of this product is not to let AI spend freely, but to set boundaries for agent payments through recharge limits, payment authorization scopes, password verification, and related confirmations.
This follows the same line as Mastercard Agent Pay: AI can participate in payments, but it must be identified, authorized, restricted, and audited. Web3 wallets face the on-chain version of the same problem. The difference is that traditional payment systems emphasize networks, merchants, issuers, and compliance responsibilities, while on-chain wallets emphasize private keys, signatures, approvals, contract calls, and user self-custody. Because of this, Agentic Wallets cannot simply copy the traditional payment path. In traditional payments, users can rely on banks, card networks, merchants, dispute handling, and risk-control systems. In the on-chain world, once a transaction is on-chain, there is often no undo button.
The central challenge is not automation, but boundaries
The more efficiency an AI agent brings, the more important the wallet becomes as the final security boundary. A future wallet should not only allow an agent to do things; it must ensure that the agent can act only within the range permitted by the user. This is why Web3 may be an especially relevant place to discuss Agentic Wallets. Discussions about AI agents often move quickly toward extreme forms of automation: AI automatically trades, manages money, finds airdrops, or performs arbitrage. Those directions may be attractive, but for wallets, the difficult issue is not automation itself. The difficult issue is a sense of boundary.
A wallet is not an ordinary application. If AI recommends the wrong song, writes the wrong paragraph, or opens the wrong page in a normal app, the problem is usually about user experience. In a wallet, if AI misunderstands one instruction, calls the wrong contract, or grants an overly broad approval, the result can become a real asset risk. The source also links this question to the extended reading titled “Sign 不只签名：当 AI Agent 替你签名，谁还握着控制权？”
Therefore, the first-principles question for an Agentic Wallet is not how many things AI can do for users. It is how users know what the AI is doing. This is also why imToken has emphasized control over the past several years. From self-custody wallets to multi-chain asset management, and now to AI co-creation and agent exploration, the continuous thread is not simply that more functions are added. The continuous thread is that users should be able to understand, confirm, and control their digital world. In the Agentic era, that principle becomes more concrete. A wallet needs to show users who the agent is, what capabilities it can call, how long the authorization lasts, whether it can operate across DApps, when confirmation is needed again, and whether the user can pause or revoke it with one action. The strength of an AI agent is that it can turn a complex process into something simple. A single user sentence can be broken into more than a dozen execution steps. That improves experience, but it also creates security challenges. The longer the path and the more intermediate links it contains, the more the wallet needs to bring key checkpoints back in front of the user. World Cup prediction activities are an event entry point, while imToken’s web agent and Discord agent are community entry points. The early form of the Agentic Wallet is already beginning in these small scenarios.