Cybersecurity – Review of Cisco’s Product Announcements at RSA Conference 2026

Cisco used the RSA Conference (RSAC) 2026 to reinforce its vision to secure the emerging agentic workforce. The provider’s portfolio enhancements spanned four strategic pillars:

In agentic identity and access security, Duo IAM now supports AI agents, enabling identity-aware access control, fine-grained permissions, and accountability mapping between agents and human owners. Integration with Cisco Secure Access adds continuous policy enforcement and intent-aware monitoring across agentic environments.

In zero-trust security for AI agents, Cisco introduced the Zero Trust AI Agent framework to establish identity-aware access, least-privilege enforcement, and secure interactions between AI agents, users, and enterprise systems. The framework reflects Cisco’s broader push toward continuous enforcement and monitoring for nondeterministic AI agents.

In AI application and workload protection, AI Defense now extends protection to AI agents, applications, and workloads through capabilities such as model vulnerability testing, policy enforcement, and runtime protection. Cisco also introduced DefenseClaw, an open-source framework designed to secure agentic applications and developer workflows.

In AI-driven security operations and enforcement, Cisco emphasized integrated visibility and distributed policy enforcement across network and security layers through Secure Access and broader platform integrations. These announcements reflect Cisco’s focus on combining identity, enforcement, monitoring, and telemetry to strengthen trust and control in AI-driven environments.

Collectively, these launches position Cisco to address key enterprise priorities around securing AI agents, enforcing continuous trust, and extending zero-trust principles to nonhuman identities and workloads. Nevertheless, enterprises will expect greater maturity in identity governance, interoperability across heterogeneous ecosystems, and operationalization of agentic IAM capabilities at scale.