Meta (META.US) Muse Goes Viral, Sparking Financial Industry Concerns; Bernstein: Bank, Insurance, and Brokerage Business Models Face Reshaping; Visa (V.US) and Mastercard (MA.US) May Be Beneficiaries

Bernstein said in its latest report that Meta's (META.US) consumer-grade AI agent Muse could become an important turning point in the consumer AI space. After launching, Muse quickly topped the U.S. app stores, has now reached 2.8 million downloads, and is beginning to be used by consumers for a range of tasks such as making appointments, ordering food, tracking flight prices, canceling subscriptions, comparing insurance products, filling out forms, and contacting customer service. Bernstein believes it is only a matter of time before other frontier AI models and tech companies follow with consumer-grade AI agents, and a "consumer AI agent race" is taking shape. This shift has already quickly been reflected in financial markets. Bernstein noted that Muse's rapid adoption has triggered investor concerns about business models that rely on consumer usage habits and switching friction. Since Muse's launch, share prices in some financial sectors, including insurance, brokerages, banks, and mortgage institutions, have fallen by a cumulative 5%-15%. However, Bernstein believes that some of the market's current "doomsday scenarios" around AI agents disrupting the financial industry are too simplistic. What truly determines whether AI agents can change the financial services industry at scale is not whether the technology can complete these tasks, but whether consumers are willing to hand financial decisions to AI, whether Financial Institutions, Inc. will allow it to access data, and whether liability attribution and regulatory rules can keep up. AI agents strike directly at the financial industry's traditional "moats"; banks, insurers, and brokerages are first in line The biggest difference between AI agents and traditional chat Siasun Robot&Automation is that they can not only provide advice, but also take action on behalf of consumers. Bernstein pointed out that an AI agent capable of automatically shopping, comparing prices, switching service providers, negotiating, and even moving funds could gradually erode a series of user habits and switching barriers that the financial industry has long benefited from, including consumers' long-term use of the same credit card, bank deposit stickiness, brokerage idle cash balances, and insurance renewal pricing. Take insurance as an example. In the past, consumers often did not proactively compare quotes from different insurers every year. If AI agents can automatically complete price comparisons and help consumers switch insurers, some of the pricing advantages insurers gain from customer renewal inertia could be weakened. Similar logic applies to banks and brokerages. AI agents can continuously seek higher-yielding cash management products, making it easier to move funds between different accounts, thereby reducing the revenue brokerages earn from clients' idle cash. For banks, easier deposit migration could push up funding costs and put pressure on net interest margins. At the same time, AI agents capable of automatically monitoring investment portfolios and conducting financial planning could also reduce consumer demand for some paid human financial advisory services. The credit card industry may also be affected. If AI can decide in real time which card to use based on rewards, interest rates, and offers, consumers' habit of long treating a certain credit card as their "go-to card" may gradually weaken. Institutions' willingness to open up becomes a key constraint Despite the huge potential impact, Bernstein believes third-party AI agents still face a core contradiction: without participation from merchants and financial service providers, it is difficult for AI agents to achieve widespread adoption; but if participation means losing the checkout entry point, customer relationships, or some economic benefits, these companies have little incentive to open up fully. Banks, brokerages, insurers, and payment companies still firmly control key links such as account login, identity verification, formal quotes, and product eligibility review. Amazon.com, Inc.'s (AMZN.US) blocking of Muse is a typical case. The report said Amazon.com, Inc. believed Muse did not clearly identify itself as an AI agent and obtained customer login credentials, while Muse has a different account of the matter. Online insurance platform Insurify also blocked Muse, on the grounds that AI automated scraping of insurance quotes may present only prices while omitting coverage limits, deductibles, discounts, eligibility conditions, and regulatory disclosure information. Bernstein expects that before the consumer AI agent ecosystem eventually matures, there will be more cases of blocking AI agents, paid data access agreements, and Financial Institutions, Inc. developing their own AI agents. There may even be a completely opposite business model: instead of AI platforms charging Financial Institutions, Inc., banks charging AI agents for data access. The report noted that Financial Institutions, Inc. still controls key gateways such as customer authentication, account access, and data-sharing permissions. JPMorgan has already begun charging data aggregators for customer data access, which provides some reference for Financial Institutions, Inc. to charge AI platforms in the future. Consumers prefer AI to "assist" rather than "decide"; trust remains a core challenge Bernstein believes the financial services industry is built on "trust," and this is precisely one of the most obvious shortcomings of current consumer AI agents. A 2026 survey of more than 2,500 U.S. consumers by TD Bank showed that 55% of respondents had already used AI to help manage finances, compared with only 10% a year earlier. However, although 62% of consumers believe AI can provide reliable information, only 18% are willing to let AI make important financial decisions independently. Consumers are more inclined to let AI improve efficiency and convenience while still keeping the final decision in human hands. Other surveys show similar results. A June survey of U.S. and U.K. consumers by ACI Worldwide and YouGov showed that only 7% of consumers are willing to let an AI assistant make purchases directly without their approval; a survey of 25,590 consumers by Accenture Plc Class A showed that 32% are willing to let AI make purchasing decisions within preset limits, but only 12% are willing to let AI make fully autonomous decisions at the payment stage. Bernstein believes that in the short term, AI agents are more likely to play the role of "assistant" rather than completely take over consumers' financial lives. AI agents move deeper into financial scenarios, while the regulatory framework still needs improvement There is another major difference between financial services and ordinary e-commerce: it is a highly regulated industry. Bernstein pointed out that AI technology is currently developing significantly faster than liability and regulatory frameworks. As AI agents begin to apply for loans, buy insurance, make payments, and even manage investments on behalf of consumers, a series of unresolved issues will become increasingly important. For example, if AI submits a loan application or purchases a financial product on behalf of a consumer, does that constitute valid authorization? How can banks confirm that AI is indeed acting on behalf of a real customer? Under what circumstances would AI-provided product optimization suggestions be regarded as regulated financial advice? If AI compares or even directly purchases insurance, does it need to obtain relevant licenses? If AI uses outdated data, acts beyond the scope of consumer authorization, or selects a product that is technically eligible but not suitable for the customer, who should ultimately bear responsibility? Bernstein expects that traditional Financial Institutions, Inc. will not passively wait for AI agents to take away customers. In addition to restricting data access and pushing for stricter regulation, many companies may launch their own AI agents, or actively cooperate with third-party AI platforms, and may even sacrifice some existing revenue to reduce the risk of customer churn. The payments industry may instead become a winner; Visa and Mastercard are favored by Bernstein It is worth noting that, unlike market concerns that AI agents will disrupt banks, insurers, and brokerages, Bernstein's view of the payments industry is clearly more optimistic. The report argues that AI agent commerce is generally seen as a potential threat to Visa (V.US) and Mastercard (MA.US), but the actual situation may be "the exact opposite." As AI agents bring more digital transactions and demand grows for value-added services such as identity authentication and risk management, payment networks may instead become beneficiaries. Bernstein believes that the early stage of AI agent commerce development may resemble the early days of e-commerce, also facing problems such as fraud, consumer confusion, and insufficient trust. What Visa and Mastercard essentially provide is a "trust network," and their advantages include not only payment itself, but also dispute resolution, transaction standards, risk management, and a vast global acceptance network. The report noted that the two companies have more than 7 billion payment credentials and cover more than 100 million merchants. Bernstein therefore believes that in the AI agent era, credit cards may still become the main payment method. As the complexity of AI agent transactions increases, payment "tokenization" may also become more widespread. Products such as Meta Muse have already begun using one-time virtual cards, while Visa and Mastercard have also launched AI agent-oriented payment solutions such as Visa Intelligent Commerce and Mastercard Agent Pay, respectively. As of September 25, Bernstein rated Visa and Mastercard "Outperform" with target prices of $450 and $710, respectively; it also rated Adyen and Affirm "Outperform" with target prices of EUR 1,600 and $110, respectively. PayPal's position is more nuanced; BNPL may occupy a favorable position The outlook for digital wallets is more complicated. Bernstein pointed out that if consumers in the future let AI agents complete payments directly, then frictions such as "guest checkout" in traditional e-commerce may no longer matter, which could weaken part of the traditional value of digital wallets. On the other hand, two-sided digital wallets such as PayPal (PYPL.US) also have the opportunity to expand their value in trust, security protection, fraud, and transaction dispute handling, while helping consumers automatically find discounts, rewards, and the optimal payment method. The report also raised that as Alphabet Inc. Class C continues to improve the Spark feature in Gemini, whether a deeper collaboration between Alphabet Inc. Class C and PayPal may emerge in the future is worth watching. At the same time, Bernstein believes that "buy now, pay later" (BNPL) platforms such as Affirm (AFRM.US) and Klarna may be in a relatively favorable position, especially in areas involving interest-bearing loans to subprime consumers, zero-interest loans, and transparency of loan terms. AI will not disrupt the financial industry overnight; trust, data, and regulation are the key Overall, Bernstein does not believe that Muse's explosive popularity means the traditional financial industry will quickly be replaced by AI agents. The report pointed out that in the early stages of development, AI agents are more likely to first solve 90% of the friction in complex processes, such as booking services, comparing insurance products, or handling travel disruptions, rather than immediately redefining everyday retail checkout or traditional financial services. Especially in the financial industry, Bernstein believes that the three major factors truly limiting the large-scale adoption of AI agents are not model capabilities, but consumer trust, data access, and regulatory rules.