Visa completes secure AI transactions: agentic commerce

What happened: Visa completed “hundreds” of secure, AI agent‑initiated transactions in production‑like tests with partners and says the work prepares for mainstream adoption in 2026. The claim appears in Visa’s December 18, 2025 press release and was reported by multiple outlets. (Visa; Financial IT)

Technical details: Trials ran over Visa’s payments rails and combined partner security tools. Akamai announced a parallel partnership to supply identity signals and fraud controls tailored for agentic commerce (AI agents acting on behalf of consumers). Coverage from Akamai and independent reporting describes the integration as focused on identity verification, real‑time signals, and fraud mitigation at checkout. (Akamai; Digital Commerce 360)

Why this matters: These tests move agentic commerce from lab experiments toward merchant pilots and infrastructure readiness. Expect vendors to expand merchant integrations, add agent‑specific identity and fraud rules, and publish developer guidance for secure agent‑initiated payments. Visa frames the effort as preparatory for broader consumer use in 2026, while partners position to supply the authentication and risk signals needed to scale.

Why It Matters

• Merchants should map where AI agents could act for customers and test agent‑specific authentication flows before pilots expand in 2026.
• Security teams must plan for real‑time identity signals and new fraud rules; Akamai’s controls serve as an early reference architecture.
• Payments architects should evaluate how agent‑initiated flows affect tokenization, liability, and chargeback processes.
• Developer teams will need SDKs and test harnesses that emulate agent behavior to validate risk controls and user experience.