To securely enable agentic commerce, organizations should adopt a reimagined security perimeter built on three capabilities: Identity-Based Access Control and Agent Personas to grant and log non-human access; Instant MCP enablement to convert APIs into secure tools without heavy development; and Real-Time protection through DLP, Bot Management, and API Security to detect and prevent agent manipulation and data exfiltration.

Many enterprises expect to deploy tailored AI agents, but only a minority currently have secure, scalable systems to govern autonomous operations, raising risks of agents acting beyond their intended scope and incomplete audit trails.

The security model should treat agents as trusted, auditable actors with defined permissions, ensuring actions are traceable and compliant while enabling scalable autonomous workflows.

Agentic commerce is expected to shift more purchasing actions from humans to autonomous AI agents by 2030, impacting consumer, enterprise, and B2B transactions.

Traditional security models are insufficient for machine-to-machine API calls, which can generate hundreds of actions per minute and lack human-like login trails.

Security should be viewed as a foundational enabler for innovation in agentic commerce, and professionals are encouraged to engage in a community for ongoing discussion.