According to PYMNTS.com, Pay3 unveiled its Agentic Payments Platform on November 4th, designed to let AI agents autonomously execute financial transactions using stablecoins. The company cited Gartner projections that 33% of enterprise software will include agentic AI capabilities within three years. Pay3’s platform integrates stablecoin payments, intelligent routing and real-time settlement across major blockchains. CEO Priya Karnik called this the intersection of “two generational technologies.” The company plans to expand using Google’s account-to-account open protocol. Meanwhile, a TRM report found stablecoins accounted for 30% of all on-chain volume but 60% of illicit activity in early 2025.
<h2 id="ai-agents-running-finance”>When Your AI Starts Spending
So we’re officially at the point where AI isn’t just recommending what to buy—it’s actually executing the transactions itself. Pay3’s platform basically lets AI systems handle everything from pricing to treasury flows autonomously. Think about that for a second. Your company’s AI could be moving millions in stablecoins while everyone’s asleep.
Here’s the thing: this makes perfect sense for enterprises drowning in routine financial operations. Cross-border payments, treasury optimization, all that boring but essential stuff. If AI can handle it faster and cheaper, why wouldn’t companies jump at that? But it raises some serious questions about oversight. Who’s liable when an AI makes a bad trade? How do you audit decisions made by algorithms you might not fully understand?
The Stablecoin Problem Nobody’s Talking About
Now let’s talk about that illicit activity statistic. Stablecoins making up 60% of illegal transactions while only being 30% of volume? That’s not just a red flag—that’s a screaming siren. Criminal actors apparently love the price stability and rapid global transfer capabilities. They get all the benefits of crypto without the volatility that makes Bitcoin impractical for actual transactions.
And we’re building entire financial infrastructure on this? Look, I get the appeal. Stablecoins combine crypto’s speed with fiat’s stability. But when you’re handing control to AI agents, the stakes get much higher. An AI doesn’t have moral qualms about where the money flows—it just optimizes for whatever parameters you set.
The Inevitable March Toward AI Commerce
Pay3 isn’t wrong about the direction we’re heading. Gartner’s 33% prediction feels conservative if anything. Once one major enterprise shows significant cost savings from AI-managed treasury operations, the floodgates will open. We’re talking about AI systems negotiating with other AI systems, settling in milliseconds, constantly optimizing cash flow.
The Google A2A protocol integration is smart—it acknowledges that these AI agents need to talk to each other across different systems. But it also creates this fascinating (and slightly terrifying) ecosystem where financial decisions happen at speeds humans can’t possibly monitor in real-time.
So where does this leave us? On one hand, we have potentially massive efficiency gains. On the other, we’re building financial infrastructure that criminals already prefer and handing increasing control to algorithms. The companies jumping into this space better have some seriously robust safeguards. Because once the AI agents start trading, there’s no putting that genie back in the bottle.
