Agentic artificial intelligence — AI systems capable of acting autonomously toward goals without requiring step-by-step human instruction — is drawing serious scrutiny from some of the most experienced voices in financial services. Among them is Chris Skinner, the veteran fintech analyst and author behind The Finanser, who argues in a newly published podcast interview that the emergence of agentic AI in payments and finance is not a recycled technology narrative, but a genuine and consequential structural shift. At a moment when the financial industry is still digesting the implications of large language models, Skinner's framing raises a harder and more urgent question: what happens when AI does not merely assist human decision-making but begins to act on its behalf?
Beyond the Buzzword Cycle
The financial industry has endured no shortage of technology revolutions that promised to remake everything and remade relatively little on the timescales originally claimed. Blockchain, the metaverse, and even early iterations of machine learning in credit scoring all passed through a predictable cycle of inflated expectation followed by quiet recalibration. Skinner, who has been writing and speaking about what he terms "agentic commerce" for some time, acknowledges that scepticism is a rational starting position. Yet in the podcast interview, he draws a clear distinction between prior waves and the current moment. The question he was pressed on — when did it click that this was real? — is itself revealing. It suggests that even among informed observers, the recognition of agentic AI as a durable force required a specific inflection point, not simply a reading of vendor roadmaps.
That inflection point matters enormously for financial institutions, regulators, and payments infrastructure operators who must make capital allocation decisions today based on technology trajectories that will not fully mature for years. Getting the timing wrong in either direction carries real cost: invest too early in immature infrastructure, and the expense is wasted; dismiss the transition too long, and incumbents find themselves architecturally unprepared when agentic systems become the dominant interface through which consumers and businesses interact with financial services.
What Agentic Commerce Actually Means for Payments
The concept of agentic commerce refers to a commercial and financial environment where AI agents — software entities operating with defined objectives and sufficient autonomy to pursue them — conduct transactions, negotiate terms, and manage financial flows on behalf of human principals. In the payments context, this is not a marginal incremental improvement. It implies that the entity initiating a payment, selecting a provider, optimising for currency conversion, or managing a subscription renewal may increasingly be a machine acting within parameters set by a human, rather than a human acting directly.
For Visa, Mastercard, and the broader payments network infrastructure, this poses a profound design challenge. Current authentication, authorisation, and fraud-detection frameworks are built around the assumption of a human actor who can be verified, prompted for a second factor, or asked to confirm an unusual transaction. Agentic intermediaries disrupt all of those assumptions simultaneously. A system that purchases cloud compute, manages supplier invoices, rebalances a corporate treasury position, and hedges foreign exchange exposure — all autonomously — requires a completely different model of identity, liability, and consent.
The Regulatory and Compliance Dimension
Regulators across major jurisdictions are already grappling with the implications. The European Banking Authority and the European Central Bank have both signalled interest in the governance of AI in financial decision-making, while the Bank for International Settlements has published research on the systemic risks of correlated AI behaviour across interconnected financial markets. The concern is not hypothetical. If large numbers of agentic systems are operating with similar optimisation objectives across payments, lending, and trading, the potential for synchronised behaviour — and the market instability that could follow — is non-trivial. Skinner's framing of agentic AI as real and structural rather than speculative gives additional weight to the urgency of regulatory engagement that goes beyond principles and into specific rule-making.
What This Means for the Industry
The practical implications for financial institutions are layered. At the infrastructure level, banks and payment processors will need to develop new credentialing and authorisation frameworks capable of distinguishing between human-initiated and agent-initiated transactions — and assigning accountability accordingly. At the product level, financial services firms will need to consider how their offerings are discovered and selected in an environment where an AI agent, not a human, may be conducting the comparison. Search engine optimisation becomes agent optimisation; customer experience design becomes agent experience design.
At the strategic level, Skinner's contention that agentic commerce is real — articulated from the position of someone who has tracked fintech evolution across multiple technology cycles — should prompt boards and executive teams to move this topic from innovation lab to core strategic planning. The institutions best positioned for this transition will be those that begin now to understand not only what agentic AI can do, but how it reshapes every assumption about who their customer is, how transactions are initiated, and what compliance and liability look like when the actor in the middle is a machine.
Written by the editorial team — independent journalism powered by Codego Press.