Visa支付网络接入ChatGPT 支持AI代用户完成购物交易

This article details the latest partnership between payment giant Visa and OpenAI, which integrates Visa's payment network into ChatGPT to launch a new AI-powered feature that completes shopping transactions on behalf of users. General readers can get a full overview of the feature's core information and practical details here.

1. Function rules: The feature is available across all merchants that accept Visa payments. After users link their Visa cards, ChatGPT can filter products based on their needs and complete the entire selection and checkout process independently. This breaks the previous limitation of similar technologies only being available at a small number of partner merchants.

2. Division of labor and security guarantees: OpenAI handles AI interaction and product selection logic, while Visa takes charge of payment authorization and fraud monitoring. Three protective mechanisms—consumption limits, mandatory transaction approval, and an allowlist of eligible merchants—have been implemented to reduce transaction risk.

3. Unspecified details: The two companies have not yet disclosed the fee rates that users and merchants will be charged. OpenAI discontinued a similar previous feature due to widespread criticism of its excessively high 4% processing fee, so the upcoming fee structure will be key to watch.

Visa and OpenAI's new cross-merchant AI-powered shopping agent represents a major new real-world implementation of AI in e-commerce, helping brands capture emerging consumer trends and align their strategic positioning.

1. Shifting consumer trends: AI agents have become active participants in economic activity, and AI-powered shopping will emerge as a new consumer shopping channel. As more users grow comfortable letting AI handle product filtering and transactions for them, brands need to adapt to the rules of this new channel in advance.

2. Cost considerations: OpenAI's previous similar feature, Instant Checkout, was discontinued after merchants rejected its 4% transaction fee as excessively high. While the new model is structured to be more merchant-friendly, fee rates have not yet been announced, so brands should monitor upcoming cost adjustments closely.

3. Entry requirements: Visa currently maintains an allowlist of eligible merchants for this new channel. Brands must meet the allowlist requirements to gain access, and should prepare in advance to address potential issues such as misaligned AI product selection and user transaction disputes.

The new AI shopping agent feature opens new growth opportunities for all types of sellers, but also requires proactive preparation to address potential risks.

1. New channel opportunity: This feature breaks the prior limitation of similar AI shopping tools that only operated on single platforms or with a small number of partner merchants. Any seller that accepts Visa payments now has the opportunity to be recommended by ChatGPT to relevant users, creating an entirely new channel for customer acquisition and transaction growth. B2B sellers can also reference Mastercard's B2B AI procurement model to serve corporate AI-powered purchasing demand.

2. Risk warnings: OpenAI's previous similar offering was rejected by merchants and discontinued over its high 4% fee. Since the new partnership has not yet announced fee rates, sellers should watch for risks of excessively high processing costs. They should also prepare for potential after-sales disputes stemming from AI product selection that does not match user needs or user claims of unauthorized transactions.

3. Entry requirements: A merchant allowlist is currently in place, so sellers need to meet the channel's requirements and prepare in advance to gain access.

The launch of the AI shopping agent model creates new business opportunities for manufacturing and processing factories, and offers new insights for factories advancing digital and e-commerce transformation.

1. Shifts in product design and development requirements: AI shopping agents filter products based on explicit user requirements such as price range and product functionality. This requires factories to clearly tag core product parameters and define price positioning during the product design and production stage, to make it easier for AI to match their products to user demand and improve the chance of being selected.

2. New business opportunities: AI shopping has opened an entirely new transaction and traffic channel. After connecting with compliant Visa merchants, factories can gain direct access to end consumers via AI recommendations to expand their sales channels. With B2B AI procurement already implemented, factories can also serve corporate AI procurement demand to grow their B2B business.

3. Transformation insights: AI + payment + e-commerce has entered the phase of real-world implementation. Factories need to accelerate their digital transformation, adapt to AI product selection rules, and seize first-mover advantage in this new track.

AI-powered shopping and payments represents a new growth direction for the e-commerce payment services industry. This partnership offers service providers insights into industry trends, customer pain points, and proven solutions to reference.

1. Industry development trends: AI agents have become active participants in economic activity, and multiple global leading companies are building out AI shopping and payment strategies. From early single-platform experiments by Amazon to Visa's current implementation covering full-spectrum C-end merchants, and small-scale B2B AI procurement already live, AI shopping and payments has moved from concept to real-world deployment, and is set to become a key industry growth avenue going forward.

2. Core current customer pain points: Key pain points already visible in the industry include excessively high fees in early models that alienated merchants, high transaction error rates, and widespread risk exposure. Banks and retail players all face concerns over overconsumption, mismatched AI product selection, and fraud and claim disputes stemming from users denying transaction authorization.

3. Referenceable solutions: The partnership uses a division-of-labor model: AI technology providers deliver interaction and decision-making services, while payment providers handle payment authorization and fraud monitoring, with additional risk controls including consumption limits, mandatory approval, and a merchant allowlist. This framework provides a proven template for all types of service providers to follow.

The Visa-OpenAI partnership offers multiple actionable insights for platform operators building out AI shopping strategies, covering platform demand, operations management, and risk mitigation.

1. Core needs of merchants and users: Both merchants and consumers want AI shopping transactions that are more convenient, lower-cost, and more secure. The failure of OpenAI's earlier Instant Checkout, which was pulled from the market due to high fees and high transaction error rates, makes clear that platforms must solve core problems of cost and security first to win market acceptance.

2. Operations and merchant onboarding reference: The division-of-labor model used here is worth adopting: AI platforms handle interaction and product selection technology, while payment platforms manage transaction infrastructure and security services, allowing both sides to leverage their core strengths. For merchant onboarding, an allowlist system to screen for compliant merchants reduces overall operational risk.

3. Risk mitigation solutions: For industry-wide concerns including overconsumption, authorization disputes, and fraud, platforms can implement controls such as consumption caps and mandatory transaction approval steps. Platforms need to build out robust risk management systems in advance to avoid large-scale claim disputes later, and set reasonable fee rates to avoid merchant pushback.

The launch of Visa's integration with ChatGPT for AI-powered agent shopping represents an important new industry development in the AI + e-commerce payment space. It demonstrates a new business model and raises novel research questions, giving it high research value.

1. New industry trends: AI has now moved beyond technology implementation to enter consumer transaction processes, with AI agents emerging as new participants in economic activity. Multiple leading companies have entered this track: from early single-platform experiments with Amazon Alexa, to Visa's deployment covering all C-end merchants, to small-scale testing of B2B AI procurement by Mastercard. AI shopping and payments is now in the early stage of commercialization, and is a key development direction for the digital economy going forward.

2. Innovative business model: This deployment uses a collaborative division-of-labor model: OpenAI leverages its AI expertise to handle user interaction, product selection, and purchase initiation, while Visa leverages its payment network优势 to manage transaction authorization and fraud risk control. This complementary model is far more market-friendly than OpenAI's earlier attempt to handle payment processing independently, and is scalable to broader use cases.

3. Open research questions: The model still faces multiple unresolved issues, including opaque fee structures, unvalidated risk control mechanisms, unclear definition of authorization liability, and ambiguous rules for fraud claims, all of which require further research to identify paths for improvement.