Wenn KI für dich zahlt: Hat die KI-Agentin das Geschäft der Cross-Border-E-Commerce-Handel beeinträchtigt?
If the era of online shopping 1.0 was the era of information yellow pages, online shopping 2.0 was mobile e - commerce, and online shopping 3.0 is social and interest - based e - commerce, then AI agentic commerce might be the biggest business revolution in the next decade.
Are you still opening ten apps to compare prices? In the future, you may not even need to move your fingers. AI will directly select the best products according to your preferences, compare prices, and automatically complete the payment. When the "shopping butler who understands your needs" in science - fiction movies becomes a reality, this business transformation driven by AI agents is quietly rewriting the rules of the e - commerce game.
AI agents may be one of the biggest variables in the future of going global for Chinese businesses.
AI breaks through geographical barriers and has cross - border insights, only to find "high - quality and low - priced" treasure products for overseas shoppers. From e - commerce platforms, independent websites to OTA and gaming & digital entertainment, for all Chinese enterprises going global involved in cross - border business, this is a strategic opportunity to seize the global consumer market.
What is AI agentic commerce?
To put it simply, AI agents actively browse products, compare prices, and place orders based on parameters set by users, such as price, brand, inventory, delivery time, and return policy. They can even complete payments with a single click in the chat box.
AI is reshaping the consumption scenario by "solving troublesome issues".
In the field of business travel, AI agents can handle everything in one go, including itinerary planning, points redemption, and automatic reimbursement, reducing the originally long - drawn - out manual operation time to just a few minutes. In the process of grabbing tickets for popular concerts, the real - time monitoring and millisecond - level order - placing capabilities of AI agents prevent consumers from missing out due to slow hands.
What's even more revolutionary is the popularization of "zero - click payment". For example, the Buy with Pro function of Perplexity enables direct transactions in the chat box, and Amazon's Buy for me service automates the entire process from recommendation to payment.
According to the data from Checkout.com's 2025 e - commerce report, 24% of Generation Z has become accustomed to letting AI dominate shopping decisions. This "silicon - based assistant" dependency is rewriting consumer psychology.
The keyword here is "zero clicks", and AI agents become the decision - makers in shopping.
There are two keywords for AI agent - based online shopping:
- User preferences, which are fed by consumers' personal data.
- Product SKUs, which are captured by AI.
The former can be fed with privacy data such as historical browsing and personal preferences, while the massive e - commerce SKU data is the core of real AI agentic commerce. Crawling this data with AI has low efficiency and accuracy.
Therefore, the data interfaces of major e - commerce platforms are the real growth engines for AI agents to exert their full power efficiently.
Will AI agents break down the data walled gardens of major platforms?
Just as Amazon benefits from the products contributed by thousands of global merchants, if AI agentic commerce can be connected to the SKU data APIs of various e - commerce platforms, allowing e - commerce websites to feed data to AI through API interfaces, using AI algorithms to recommend the best products to consumers will no longer be just an idea on a PPT.
The premise is whether the platforms are willing to contribute such data interfaces. The advantage is that they can embrace AI agents as a new traffic entry, while the disadvantage is that they may lose direct interaction with users and brand loyalty. It's a double - edged sword.
The core of e - commerce always lies in the user interaction entry.
Similar to the situation in China, the data of foreign e - commerce platforms has never been truly interconnected. Their core strategy is to build "walled gardens". Take Chinese e - commerce platforms like Taobao and Xiaohongshu as examples. Each has its own logic for product promotion. Just as people's joys and sorrows are not mutually understood, the algorithmic logics of different platforms are not interoperable. Even if there is some data cooperation between the front - end and back - end of different platforms, it usually only establishes a data handshake area, and both sides still maintain core barriers and retain their own data sovereignty.
The revolution of AI agents lies in the possibility of removing the walls of the data garden and blurring the boundaries of platform data sovereignty.
It's not hard to imagine that in the future, the cookie details when users visit web - based pages and the privacy permissions when installing apps will all be related to the data access permissions of AI agents.
The "power shift" in the e - commerce ecosystem and the dilemma of payment security
When 76% of global AI users start to rely on chatbots for shopping advice, and when 50% of German consumers consult AI rather than search engines first - this "silent revolution" is shaking the foundation of the e - commerce ecosystem.
The "walled garden" model of traditional e - commerce platforms is facing collapse, and the iterative path of payment interfaces is also clear - from the web - based to the mobile - based, and now to AI agents.
Each evolution reduces the participation of "humans". If e - commerce platforms, independent websites, etc. stick to their closed data systems, they will face the risk of being "ignored" by AI agents.
However, the cornerstone of the payment field is still trust and security. As AI agents get more involved, the payment industry is experiencing its third paradigm shift after mobile - based payment.
Therefore, payment authenticity verification has become an important infrastructure in the era of AI agentic commerce.
In the cross - border payment podcast "Endless Payments", Steven, the head of customer solutions for Mastercard in Greater China, and Checkout.com, an acquirer, pointed out that there are three levels of verification when talking about identity verification:
- Verify whether it is the consumer himself/herself behind the agent.
- Verify whether the payment token is being used by a real agent.
- Verify whether the transaction order is correct.
Meanwhile, payment companies play a key role in connecting merchants and AI platforms - evolving from simply processing payment transactions to building a secure, trustworthy, and intelligent collaborative ecosystem.
AI agents usually access merchant websites in a "invisible" way.
What does "invisible" mean?
- Actions are imperceptible.
- There is no data interaction.
- The payment is not made by a real person.
This will lead to three major transformations:
- From context - lacking shopping to merchant - empowered contextual transactions
- From one - way operations to guided interactions constrained by security guards
- From a generalized experience to dynamic personalized services
To achieve this vision, the first batch of global payment companies have started to jointly promote infrastructure construction with card associations. The general process is as follows:
1. Agent identity registration and identification
From the traditional KYC (know your customer) to the latest KYA (know your agent). Create a unique digital identity ("digital fingerprint") for each AI agent.
- Purpose: Enable all parties involved in the transaction (merchants, banks, card associations) to clearly identify that the transaction is initiated by an agent rather than a real person. This is the foundation for building trust.
2. Special payment credentials for AI agents
Allocate suitable dedicated payment tools (such as special tokens and virtual cards) for AI agents instead of letting them use consumers' personal credit cards, ensuring transaction security, scalability, and enabling financial institutions to effectively manage and control risks of agent transactions.
- AI agent - specific network tokens: Tokens issued with unique identifiers and rules can clearly distinguish that the transaction is initiated by an AI agent (rather than a human).
- Virtual cards: They still have practical value in specific scenarios but need to be integrated as part of a managed framework rather than as scattered ad - hoc solutions.
- Credentials based on digital wallets: Allow AI agents to conduct transactions through digital wallets and enjoy the same level of security and transparency as human transactions.
3. Identity verification
Confirm the identity of the end - user as the consumer himself/herself authorizing the transaction through secure verification methods (such as biometrics), while ensuring that the consumer's sensitive payment information will not be leaked to the AI agent.
4. Empowering personalized decision - making
With the consumer's authorization, provide the AI agent with consumption preference data (such as "frequently bought coffee"), enabling the AI to make smarter and more personalized shopping recommendations (such as "recommend a hotel near a coffee shop") and improving the user experience.
5. Managing authorization and having full control
To ensure that the AI agent acts within the scope defined by the consumer and to provide authorization credentials for possible order disputes in the future, protecting the rights and interests of all parties, consumers can pre - set the operating rules (such as budget, conditions, and time) for the AI agent.
Merchants' "AI - friendly" transformation and ecosystem co - construction
The rise of AI agents will bring three historical transformations to the current e - commerce ecosystem:
1. Power shift of traffic entry:
From SEO in the PC era, to algorithmic recommendation in the mobile era, and now to the reconstruction of the "people - product - scenario" by agents in the AI era, brands and platforms must rethink how to make product information "understood" and recommended by AI. Traditional psychological marketing is being replaced by algorithmic problems. Structured data and open APIs are the keys for merchants to connect with AI agents.
2. Reshaping of business forms:
AI agents are not affected by advertisements and make decisions more based on hard indicators (such as price, inventory, and return policy). The space for brand stories is compressed, and the shopping process becomes more rational and automated. Enterprises need to find a balance between "automation" and "personalized services", build "agent - friendly platforms", and continuously innovate payment solutions to adapt to the AI transformation.
3. Upgrade of payment security and trust:
AI agent payment (Agent Pay) is promoting the evolution of payment technology: Payment tokenization, identity verification, biometrics and other technologies should be transplanted to the AI agent scenario to ensure transaction security. The new "Know Your Agent (KYA)" framework guarantees the identity and permissions of agents in each payment link, reducing the risk of fraud and enabling consumers to always have control over their fund authorization.
Future development roadmap
Although AI agentic commerce is in its infancy, its evolution path is clear:
The future is here. From "science - fiction imagination" to "business implementation", AI agent payment is rewriting the underlying logic of consumption and payment.
AI agentic commerce actually brings significant benefits to Chinese cross - border e - commerce. AI agents are borderless and can select "high - quality and low - priced" products that best suit the personal preferences of global consumers.
For e - commerce platforms, independent websites, OTA platforms, gaming & digital entertainment, and even fintech platforms, new business models and paradigms are emerging.
This article is from the WeChat official account "Brand Factory BrandsFactory" and is published by 36Kr with authorization.