-+ 0.00%
-+ 0.00%
-+ 0.00%

AI agents are setting off a new battle in the payment industry! Bernstein: Visa (V.US), Mastercard (MA.US), etc. welcome new opportunities

Zhitongcaijing·10/14/2026 07:09:05
Listen to the news

The Zhitong Finance App learned that investment bank Bernstein pointed out in a research report released recently that Muse launched by Meta (META.US) has opened a new era of artificial intelligence (AI) intelligence for consumers and a new round of competition. However, the bank believes that as the commercialization of smart devices evolves, payments may become more strategic, and the next major war in the payment industry may begin.

Bernstein pointed out that with the rapid evolution of commercialization of smart devices, payments are also keeping up. Muse is already being used to perform a variety of tasks: shopping (Meta says this is quickly becoming one of its biggest drivers of use), tracking prices after booking flights and hotels, making reservations, cleaning inboxes, and matching clothes.

However, the smart payment experience is still a bit clumsy. When checking out, Muse will use Stripe's Link or Shop Pay to fill in the merchant's own checkout page. AI intelligence startup Instinct said 40% of its users would share personal credit cards within three weeks, but its early users were probably more tech-oriented and more profitable than ordinary consumers.

The fundamental question surrounding consumer AI agents in the payment industry is, in an intelligent world, who will control the payment, liability, and dispute resolution framework? What role will payment networks and PSPs (payment service providers such as Stripe, Adyen, etc.) play in this? Will this create incremental revenue opportunities (e.g. for V/MA, Adyen, Stripe), or if payments are abstracted or tokenized, putting some payment processors at risk of commercialization?

Trust remains a major bottleneck in intelligent commercialization — only 23% of US consumers say they trust generative AI to process payments instead of them. The traditional business model is designed around “people in the ring.” In a world where smart bodies crawl information, fill out forms, and buy products on the internet, there are a series of issues surrounding trust: How do people know that an intelligent body is a “good intelligence”? How can transaction programmability be achieved? How can I verify my identity (such as two-factor authentication in Europe, or 2FA)? How do you ensure user consent and user intent? How is the accountability framework designed? How are disputes handled? How to reduce the risk of fraud? How to manage chargebacks? How to set up an audit trail? How to optimize different payment methods? How can personalization be achieved? And how to address the fragmentation of standards and technology?

The battle for control over smart transactions

A “tug-of-war” is about to begin over control of smart transactions and control of certificates/payment tokens.

Smart devices bring far more complexity to a transaction than in the past. In the age of intelligent commerce, trust and governance are critical (even more important than e-commerce), and extremely challenging — and entities that can provide intelligent certificates/tokens and associated governance mechanisms will reap huge benefits.

To be fair, intelligent commerce still accounts for very little in e-commerce, but it is developing rapidly, and payment companies are preparing for the intelligent world.

Both major payment networks, Visa (V.US) and Mastercard (MA.US), have launched smart tokens (Visa Intelligent Commerce, Mastercard Agent Pay). More data and contextual information (such as deeper insight into transaction types, where the tokens are used, and the identity of the payer) have been added to the token certificate, while adding a guarantee mechanism (signals based on behavioral history and related configurations).

Currently, it is still unclear how these products are being used and whether card issuers are actively involved (for example, around smart device control and accountability frameworks). In any case, 100% of smart transactions will be tokenized (but not necessarily using agent-specific tokens). Meanwhile, Stripe launched Secured Payment Tokens (SPT) for smart transactions a year ago.

It's important to note that currently, the Muse or Instinct service probably doesn't use Visa and Mastercard's “smart exclusive” tokens, but instead uses Link or one-time virtual cards — which means that payment networks can still obtain these transactions, but are unable to obtain the lucrative “smart” token business, and the latter may become a platform for distributing smart-oriented services in the future. Muse transactions are also covered by Link's purchase protection mechanism.

Interestingly, Visa indicated in July that it had partnered with Meta to use smart tokens within Facebook and Instagram, but did not find any specific information specific to Muse.

Whether payment networks can seize a leading position in smart transactions through “smart tokens” may ultimately depend on two words — liability shift (liability shift). Even if other tokens are just as technologically advanced, or even more advanced, payment networks still play a key role in dispute handling and accountability frameworks, and this may become extremely important in the field of intelligent commerce.

Learn from history?

The battle for control over smart transactions is similar to the 2014 token war. Twelve years ago, along with the launch of Apple Pay, Visa and Mastercard also launched their own token services. At the time, there was also a lot of competition over who would control the token — competition between payment networks, card issuers, and PSP.

Today, despite the coexistence of multiple tokens, payment network tokens have been widely adopted and currently account for 60% of e-commerce transactions. Tokenization also provides a real leading position for payment networks, enabling them to distribute their own services through this channel, such as providing fraud scores, automatically updating payment vouchers, and push provisioning (push provisioning) of card issuer certificates to merchants. Today, they are finally starting to price these services, and there is still plenty of room for pricing growth in the future.

The potential market space is likely to be huge

The potential benefits are likely to be significant. For example, Stripe currently charges 0.15 cents for SPT. Once payment networks' smart tokens are widely adopted, they can distribute their services through these tokens. Under an optimistic scenario, if smart transactions can eventually account for a single-digit to high-single-digit percentage of overall consumer transactions, this may contribute several percentage points to the growth of payment networks.

What's worth watching?

Factors to focus on include: system integration between card issuers and merchants (rather than just being enabled by default); holiday shopping season (and smart device adoption); new platform partnerships; whether Google or Meta's Muse will implement smart tokens; and smart experiences launched by banks (such as Amex and Chase).

Incidentally (which is also indicative for platforms), none of these recently launched smart products currently allow smart devices to control merchant payments (or assume the role of merchant recorder, that is, Merchant of Record) — this model is quite complicated, to say the least. Instead, these products mainly provide consumers with a frictionless way to complete tasks and make purchases (such as Muse to complete payments via Link or a one-time virtual card).

There are serious difficulties in integrating merchant systems, and mastering the usually merchant's own checkout experience also faces many challenges, including managing conversion rates, real-time pricing, post-purchase services, synchronizing merchant product catalogs and inventory, applying member loyalty and discounts, fulfillment, and multiple shopping baskets.

Investment impact

Bernstein believes that payments are beneficiaries rather than losers, and in a world of consumer AI agents, payments may become more strategic.

One possible scenario is that in the future, consumers will make more transactions through smart devices (smaller shopping baskets and more long-tail products discovered), which will benefit payment companies, including bank card payment networks.

In addition, smart devices can also further increase the online penetration rate of complex services and some retail categories. This may be less intuitive, but increased e-commerce penetration is generally very beneficial to bank cards because the online penetration rate of bank cards (and the additional rate of value-added services, or VAS) is higher.

Consumer AI agents may also mean more complexity in smart authentication/verification, consumer intent, dispute handling, and agent-specific cards/credits — thus creating opportunities for V/MA and payment processors (such as Stripe, Adyen, Checkout) to differentiate and provide value-added services.

It's important to note that 100% of smart transactions will be tokenized (this means V/MA may earn more revenue per transaction, while also distributing more service channels for it). Agent-to-agent (agent-to-agent) commercial transactions may also follow, which may require a richer set of services, and therefore also form a favorable factor.

Bernstein has been thinking deeply about this issue, and believes that digital wallets (such as PayPal) are in an interesting position in the field of consumer AI smart devices and have strategic options (for example, if it is difficult for smart agents to find effective commercialization models).

On the one hand, if smart bodies are responsible for processing people's payments — then do people really care about the friction caused by “guest checkouts (guest checkouts)”? On the other hand, bilateral wallets like PayPal (PYPL.US) can significantly expand their value proposition, including trust, security, fraud prevention, and dispute resolution, while also helping consumers dynamically find cost-saving opportunities, rewards, and personalized services, and optimize payment methods (such as choosing between BNPL and credit/debit cards).

Buy-To-Pay Later (BNPL) providers such as Affirm (AFRM.US) and Klarna (KLAR.US) may also be in an advantageous position, particularly in the following areas: interest-bearing loans (including costs such as late fees) for near-prime (near-prime) and subprime (subprime) customers, dynamic credit (can they be negotiated with agents?) , 0% APR loans (for example, for quality customers), and transparency of terms.

Bernstein gave Visa, Mastercard, Affirm, Adyen (ADYEY.US), Block (XYZ.US), and Toast (TOST.US) “outperform the market”; PayPal, FIS (FIS.US), Fiserv (FISV.US), Global Payments (GPN.US), and Klarna were rated “on par with the market.”

21.png