What are experts saying the 'future of payments' will look like in an AI-focused world?

Experts predict a fundamental restructuring of commerce and payments driven by AI agents that autonomously execute transactions on behalf of users . This shift to an "agentic economy" will move beyond AI-assisted recommendations to end-to-end purchasing without direct human intervention [6, 17]. Projections for this transition are aggressive, with some predicting AI agents will execute the **majority of global payment volume within 10 years** [11, 18] and soon account for the majority of all internet transactions . This paradigm shift will redefine consumer interaction with financial services, potentially making AI applications the primary interface for complex transactions like mortgages within the next two years . However, this evolution raises significant concerns for retailers, who fear losing their direct connection to customers and their status as the merchant of record as AI intermediaries make purchasing decisions .

The existing financial infrastructure is widely seen as incapable of supporting the scale and nature of an agentic economy . The current system's reliance on traditional banking hours and slower payment rails cannot accommodate the millions or billions of AI agents that will need to transact globally, instantly, and programmatically on a 24/7 basis [7, 10]. Furthermore, providing AI agents with direct access to conventional bank accounts or credit card details is considered infeasible due to security and control risks . This technological gap necessitates the development of new, software-native financial rails designed for machine-to-machine interaction, with some analysts projecting a need for blockchains that can support **over a million transactions per second** to handle the anticipated volume [7, 19]. The future of finance will likely see companies themselves evolve into software-based entities where core functions like contracts and governance are executed by smart contracts interacting with AI [5, 29].

To address these infrastructure challenges, many experts point to stablecoins and blockchain technology as the foundational payment layer for AI agents [6, 7]. Stablecoins like USDC, which are presented as fully-reserved digital dollars backed by liquid assets, offer a trustworthy and efficient medium of exchange for the internet economy, avoiding the volatility of other cryptocurrencies [4, 14]. This model suggests a future convergence where digital wallets holding stablecoins will either coexist with or potentially replace traditional bank accounts [21, 22]. Crypto wallets and stablecoins are viewed as the most suitable "financial primitive" for AI agents, providing a secure and programmable way for them to transact on behalf of users without compromising legacy financial accounts .

While the vision for an AI-driven payment future is transformative, the current application of AI in finance presents a more complicated reality. Presently, one of the most significant use cases for AI in the financial sector is enabling sophisticated fraud, a problem that is **growing 18-20% annually** [1, 23, 28]. In parallel, AI offers a massive opportunity to automate inefficient, labor-based back-office processes that constitute a large portion of the financial services market [1, 24]. The transition to agentic commerce also introduces unresolved questions regarding legal liability, as it remains unclear who is responsible—the consumer, the bank, or the AI developer—when an autonomous agent executes a faulty transaction . This tension between AI's potential for fraud and its promise of efficiency, coupled with unresolved legal frameworks, highlights the significant hurdles that must be overcome to realize the predicted future of payments.