Cisco has launched a suite of security innovations designed to protect the emerging agentic artificial intelligence (AI) ecosystem, where software acts autonomously rather than simply responding to queries.
The announcements, made at RSA Conference 2026 on March 24, aim to remove key barriers to enterprise adoption of AI agents by embedding security at every stage of deployment.
Cisco’s new solutions focus on three pillars: Protecting the world from AI agents, protecting agents from external threats, and detecting and responding to incidents at machine speed.
The company is extending Zero Trust Access to AI agents, enabling organisations to register agents, map them to accountable human owners and enforce strict, task-specific permissions.
To secure agents before deployment, Cisco introduced AI Defense: Explorer Edition, allowing teams to red-team models, test for vulnerabilities such as prompt injections and jailbreaks, and generate actionable security reports. The Agent Runtime SDK further embeds policy enforcement into agent workflows across major AI platforms including AWS, Google, and Azure.
Cisco also unveiled DefenseClaw, an open-source secure agent framework integrating tools like Skills Scanner, MCP Scanner, AI BoM and CodeGuard to ensure safe, sandboxed agent deployment. For runtime protection, it collaborates with NVIDIA’s OpenShell to automate security checks and maintain zero-trust integrity.
On the detection and response front, Cisco is expanding the Agentic SOC, integrating AI-driven capabilities into Splunk Enterprise Security. New tools such as Exposure Analytics, Detection Studio, and specialised AI agents — including SOP, Triage and Malware Threat Reversing Agents — allow SOC teams to detect and mitigate threats at machine speed while reducing analyst fatigue.
“AI agents are a new workforce that expands organisational potential, but security is key to trust and adoption,” said Jeetu Patel, Cisco President and Chief Product Officer.
The platform is designed to help enterprises move from pilot to production with confidence, ensuring agents are tested, hardened and monitored across their lifecycle.
The rollout underscores Cisco’s commitment to addressing AI security risks proactively, enabling businesses to scale agentic workforces safely amid rising enterprise adoption.
