In a recent LinkedIn post, Linas Beliūnas discusses a significant advancement in artificial intelligence: the ability for AI agents to autonomously create and use payment methods to transact in the real economy. Beliūnas highlights a demonstration where Claude AI generated a virtual Visa card and used it for online purchases, a capability he argues is a preview of how AI agents will operate.
AI Agents Move Beyond Research to Transaction
Traditionally, AI agents have excelled at tasks like research, planning, coding, and navigating websites. However, Beliūnas points out that they have typically stalled at the checkout page, unable to complete purchases. This new capability, demonstrated through tools like AgentCard which integrates with Stripe, bridges that gap.
“Claude can instantly create a prepaid Visa card with a fixed limit (e.g. $10, $25, $100) funded via Stripe. The agent uses that card to buy things online – API credits, domains, SaaS tools – without ever seeing your real banking details.”
As Linas Beliūnas explains, these generated cards are designed to be single-purpose and disposable, mitigating risk. This approach allows AI agents to execute full workflows, from research to deployment, without human intervention for payment details.
The Rise of Trust Infrastructure for the Agent Economy
While the ability for AI agents to transact at machine speed opens up new possibilities, Beliūnas also raises critical concerns about the potential for scaled fraud. He emphasizes that the defense against such automated fraudulent activity must also be AI-driven.
“Because in a world where agents can spin up thousands of transactions, the defense layer has to be AI vs AI.”
According to Linas Beliūnas, platforms that offer real-time AI risk decisioning, such as Oscilar, are becoming essential infrastructure. He posits that the future of the agent economy hinges not just on intelligence, but on robust trust mechanisms.
Enabling Autonomous Workflows
Beliūnas’s analysis suggests that the integration of payment capabilities transforms AI agents from tools that merely assist in tasks to autonomous actors capable of completing entire business processes. This includes deciding, purchasing, and deploying resources.
“The agent economy won’t run on intelligence alone. It will run on trust infrastructure.”
In Linas Beliūnas’s view, this evolution necessitates a new layer of security and trust. The combination of payment rails like Visa enabling transactions and platforms like Oscilar ensuring system integrity is crucial for the sustainable growth of an economy driven by AI agents.
📝 About This Content
This article is based on insights shared by Linas Beliūnas on LinkedIn.
📅 Originally posted on March 13, 2026 | View original post on LinkedIn →