• Plaid CEO predicts AI platforms will monetize financial services through embedded payments
• Card issuance could become an early revenue opportunity for ChatGPT and Claude
• Banks face growing competition for control of customer relationships
• Fitch identifies moderate AI disruption risks, particularly for retail and SME banking
• Financial institutions expected to develop their own agentic AI services
• Plaid warns AI agent swarms could create unprecedented banking security threats
• US fraud losses could exceed projected $40 billion by 2027
• Industrywide cooperation essential to combating AI-powered fraud

Artificial intelligence platforms could become major financial services providers, challenging banks' control over customer relationships while exposing the industry to unprecedented fraud risks, according to Plaid co-founder and CEO Zach Perret.

Speaking at the Money20/20 conference in Las Vegas, Perret predicted that financial services would become "a monetization lever" for AI platforms such as OpenAI's ChatGPT and Anthropic's Claude as their capabilities expand beyond providing information to executing financial transactions.

He identified payment card issuance as an obvious starting point, with AI companies potentially generating interchange revenue from cards embedded within their platforms.

"That's a pretty straightforward thing to have happen," Perret said. "Maybe they do more lending products and so forth, but I think the card issuing will make a lot of sense in step one."

The development raises fundamental questions about whether consumers will continue managing their finances through traditional banking applications or increasingly delegate financial decisions to AI agents.

Perret questioned whether customers' entire financial lives could eventually reside within AI platforms, which he described as "harnesses," or whether banks and fintech companies would develop competing agentic services.

The issue has become a central industry concern as financial institutions assess the implications of AI agents acting autonomously on customers' behalf.

Fitch Ratings recently characterized AI-related disruption as a moderate risk to banks over the next five years, particularly through increased technological competition affecting financial products and processes.

Retail banks and institutions serving small and midsize businesses could face the greatest exposure because their standardized, high-volume products are relatively easy to replicate.

However, Fitch emphasized that the extent of disruption would depend on how quickly consumers embraced AI-led financial services.

Banks are already considering defensive strategies. Fifth Third CEO Tim Spence recently suggested that established institutions could counter external AI platforms by incorporating similar capabilities into their own applications while emphasizing bank-grade information security.

Perret expects financial institutions to develop increasingly sophisticated AI agents tailored to their customers' requirements.

Mortgage providers, for example, may prefer to retain conversations with prospective borrowers within their own platforms rather than surrender those interactions to external AI services.

"I would want the mortgage conversation actually to happen in my four walls, as opposed to outside," he said.

He predicted another generation of AI-powered banking assistants, although he acknowledged that earlier fintech chatbots had largely failed to deliver meaningful customer value.

Rather than one platform dominating every financial interaction, Perret anticipates a fragmented market in which consumers use different services for different purposes.

Routine activities, such as transferring money between checking and savings accounts, could migrate toward AI platforms, while more complex transactions, including mortgages, might remain within banking applications.

However, the opportunities presented by agentic finance are accompanied by potentially severe security threats.

Perret warned that Deloitte's projection of US fraud losses reaching $40 billion by 2027 could underestimate the scale of the emerging problem.

"That's probably not including the risk of agent swarms attacking banks, which is even more terrifying to me, frankly, than AI-enabled fraud or deepfakes or things like that," he said.

The prospect of coordinated autonomous agents targeting financial institutions has intensified concerns about the ability of conventional fraud detection systems to withstand increasingly sophisticated attacks.

Perret acknowledged that discussions within Plaid's leadership team about emerging security threats had become deeply concerning.

"We are rapidly doing everything we can from a security perspective, from a fraud perspective for our customers," he said, warning that the industry faced a potentially frightening year.

Plaid, which connects financial accounts with external applications and services across approximately 12,000 institutions, is developing fraud prevention products using its proprietary data network while collaborating with other technology providers.

"We're just in a dead sprint using AI to fight AI fraud," Perret said.

He emphasized that combating emerging threats would require industrywide cooperation rather than isolated institutional responses.

"Fraud is a collaborative game," he said, warning that financial institutions could face sustained security challenges for several years as increasingly capable AI systems transform both legitimate financial services and criminal activity.