FIS Global Calls on Banks to Authenticate AI Agent Authority in Payments
Originally published:
FIS Global, identified as the first major payments processor to publicly address AI agent authorization, is working with financial institutions to apply differentiated fraud scoring and track agentic transactions across the payment lifecycle. The firm proposes purpose-built virtual cards with restrictions on merchant, amount, and time window as a near-term control layer. A phased approach keeps humans involved while intent-capture…
Why it matters for finance
Treasury and AP teams deploying AI agents for purchasing or payment execution need to understand emerging authentication gaps and evaluate virtual card controls to manage authorization risk.
- Map all AI agents that currently hold or invoke stored payment credentials within your organization's payment stack.
- Engage your payments processor — beginning with FIS Global if already a client — to determine whether agentic transaction identification and differentiated fraud scoring are available or on the near-term roadmap.
- Replace general-purpose credentials used by AI agents with purpose-built virtual cards that carry merchant-specific, amount, and time-window restrictions.
- Establish a mandate definition process so each agent's authorized scope is documented and can be passed as an intent signal through the payment lifecycle.
- Retain human approval checkpoints for agent-initiated transactions above defined thresholds while intent-capture infrastructure is still being built out.
- Monitor evolving industry frameworks — including network-level Credential Convergence and issuer-level proof-of-intent schemes — to align internal controls as standards converge.