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, allowing AI systems to manage pricing, billing and treasury flows. CEO Priya Karnik positioned this as the intersection of “two generational technologies” – agentic AI and stablecoin payments. The company plans to expand capabilities using Google’s new account-to-account open protocol, and supports enterprise use cases including cross-border payments, treasury optimization, and stablecoin acceptance and issuance.
When AI Starts Handling the Money
Okay, so now we’re giving AI agents the keys to the stablecoin treasury. That’s… a bold move. Pay3’s vision of “AI-native commerce” sounds futuristic and efficient – who wouldn’t want automated systems handling cross-border payments and treasury optimization? But here’s the thing: we’re barely regulating human financial institutions properly, and now we’re handing decision-making authority to algorithms that can make million-dollar mistakes in milliseconds.
I keep thinking about the Gartner prediction – 33% of enterprise software having agentic AI capabilities in three years. That’s an incredibly fast adoption curve for technology that’s essentially autonomous money managers. And we all remember how well automated trading algorithms worked out during flash crashes, right? The difference now is we’re talking about systems that can move real value across borders without human intervention.
The Stablecoin Problem Nobody’s Talking About
Now for the really concerning part. That same PYMNTS report drops this bombshell: stablecoins accounted for roughly 30% of all on-chain volume in early 2025 but made up 60% of illicit volume. Let that sink in. Criminal actors are apparently twice as likely to use stablecoins for illegal activities compared to other cryptocurrencies.
Why? Because they combine “price stability, rapid global transfer, and blockchain-native transparency combined with anonymity.” Basically, criminals get all the benefits of crypto without the volatility. And we’re building systems where AI agents can autonomously move these exact assets around? That seems like asking for trouble. The report notes that stablecoins’ share of illicit activity is “materially larger than their share of overall volume” – which should be a massive red flag for anyone building financial infrastructure.
The Future of Finance or Regulatory Nightmare?
Look, I get the appeal. Autonomous AI agents managing corporate treasuries could revolutionize efficiency. But we’re creating systems where algorithms can move value across jurisdictions instantly, using instruments that are disproportionately favored by criminals, with minimal human oversight. What could possibly go wrong?
The institutionalization of stablecoins is indeed underway, as traditional financial institutions explore issuance and integration. But we’re building this future on shaky foundations. Pay3 talks about making finance “smarter, faster and more accessible” – all great goals. But without corresponding advances in security, regulation, and oversight, we might just be building the perfect system for financial crime at scale.
So while the technology is undeniably cool, maybe we should pump the brakes just a bit. Because once these AI agents start moving real money autonomously, there’s no undo button when things go wrong.
