AI agents necessitate new security standards for agentic payments

The rise of AI-driven agentic payments introduces new challenges for IT security and digital identity. Unlike traditional online payments where a person directly authenticates a transaction, AI agents act on behalf of users, selecting products and potentially triggering payments autonomously.

To ensure security, technical infrastructures must verify not only the identity of the human user but also the specific permissions granted to the digital agent. This requires a tighter integration of authentication and authorization, allowing users to set constraints such as price limits, specific merchants, or product categories. Mastercard and Nickel have conducted pilot tests for authenticated agent-based transactions under real-world conditions to address these requirements.