AI agentic payments
AI agentic payments promise a continuous, 24/7 financial ecosystem operating at maximum economic efficiency, but taking the human out of the loop exposes a gaping regulatory and insurance chasm. Matthew Challis explores
Reducing our workload
For most of human history, in one way or another, we have dedicated swathes of time and energy to reducing our workload. From the invention of the wheel to the introduction of Microsoft Teams, humans have strived to take a little bit off their plates with each new technological advancement.
AI’s vast and unprecedented adoption over the past five years has, in no small part, reshaped institutional finance; the technology has transitioned from the rigid, rule-based automations of yesteryear, developing into one comprising advanced neural networks and capabilities. With the increasingly prevalent shift towards a ‘digitally native’ financial world, one of the latest trends arrives in the form of AI agentic payments, wherein AI agents autonomously discover, negotiate, and even execute purchases on behalf of firms.
The digital assets space
DeFi markets are defined by their programmability and 24/7 nature, which are frequently cited as the most desirable aspects of the digital assets space. Traditional cross border B2B payments are far from frictionless. Operating delays and the constraints of TradFi working hours can result in a myriad of issues for any party at any stage of a transaction – from approval setbacks to interventions around execution.
What AI agentic payments set out to do, according to AI-powered financial operations platform Ramp, is sit somewhere between an automatic payment and a human-judged one, combining the former with the latter.
It comes as little surprise, therefore, that, with such a promising and capable technology, B2B fintech platform Galileo projects that, by 2032, the agentic payment will grow from US$7 billion to US$93 billion, and Bain & Company, a management consultancy, estimates that the US agentic commerce market could reach US$300–500 billion by 2030, representing 15–20 per cent of domestic e-commerce volume.
Without human intervention
But without human intervention, how can institutions rest assured that their money is in safe hands? What about the decrease in security? Are the AI agents even capable of human judgment?
Most importantly, what happens when something invariably goes south, and who or what is to blame when that happens?
