When an agent uses a digital service…
At the heart of the prototype is a simple scenario: an autonomous agent uses a paid digital service. The agent calls an AI service and receives a machine-readable payment request in return. It then processes the necessary information, authorizes the payment via a wallet, and, after successful verification, gains access to the desired service. The entire process takes place without human intervention, based on the defined rules and permissions. At the same time, the prototype also considers the service provider’s perspective. The provider protects its service with a machine-readable payment request and only grants access after successful verification. This makes it clear what requirements arise not only on the side of Machine Customers, but also for providers of digital services.
… standards become crucial
A key insight from the prototype is the importance of shared standards. For autonomous systems to interact with one another, payment terms, authorizations, and communication processes must be described in a machine-readable and standardized way. Without shared standards, autonomous transactions can neither be efficiently scaled nor operated in a controlled manner. Especially for banks and insurance companies, interoperability thus becomes a central prerequisite for future machine-to-machine interactions. In the prototype, this is examined exemplarily using the standardized payment protocol (x402), which integrates payment information directly into the interaction between agent and service. At the same time, it becomes apparent that such approaches must be carefully secured. For example, a payment request must not be reusable or detachable from its original context. It must always be clear what an agent is paying for, who the payment goes to, and under what conditions it was released. Precisely for this reason, robust rules and control mechanisms are needed before such interactions can be widely deployed.
Blockchain as a possible enabler
The prototype uses blockchain technology as the technical foundation for payment processing and verification. However, the focus is not on the blockchain itself. Rather, what matters is the question of how transactions between autonomous actors can be handled in a traceable, verifiable, and trustworthy way. In this context, the blockchain serves as a possible settlement layer for programmable payments and verifiable transactions. Regardless of the infrastructure, the prototype shows that many of the necessary technological building blocks are already available today. At the same time, it makes clear that Machine Customers raise new questions. Who authenticates an autonomous agent? How are permissions managed? How can decisions be traced and regulatory requirements met? And also: How do autonomous systems build trust? Particularly in regulated industries such as banking and insurance, it quickly becomes obvious that a purely technical perspective is not sufficient. The technologies must be integrated into existing ecosystems in a controlled, interoperable, and regulatorily compliant manner. Governance, compliance, and auditability will be crucial for this.
