#USA #San Francisco #Cisco #Zero Trust #AI Defense

Cisco's Innovative Security Approach for the Agentic Workforce

At the RSA Conference 2026 in San Francisco, Cisco unveiled major advancements in security tailored for the agentic AI ecosystem. As organizations increasingly rely on AI agents for various tasks, it's crucial to address associated security challenges. Cisco's new solutions focus on Zero Trust principles, ensuring that AI agents operate safely and effectively.

Zero Trust Access for AI Agents

A key highlight of the initiative was the extension of Zero Trust Access to AI agents. This development enables organizations to verify each agent's identity, establish accountability, and secure agent actions. With new Identity and Access Management (IAM) capabilities from Cisco Duo, businesses can now register AI agents, map them to responsible human owners, and maintain visibility over agent activity. This approach is vital in today's landscape, where malicious actors increasingly target identity authentication components.

Enhanced Identity Management

Cisco's expanded Duo IAM features allow businesses to:

• - Register AI Agents: Ensure every AI agent has a verified identity, which aids in tracking actions and establishing accountability.
• - Gain Tool Visibility: Discover both AI and non-human identities within the organization to assess existing AI usage accurately.
• - Implement Strict Access Control: Assign permissions tailored to specific tasks, routing traffic through a secure gateway to eliminate potential security blind spots.

As Jeremy Nelson, CISO for North America at Insight, pointed out, organizations are excited to adopt AI but require solid security mechanisms to avoid gaps that could jeopardize data integrity.

Safeguarding the Agentic Workforce

Cisco does not stop at safeguarding organizations from agents; it also aims to protect agents from external threats.

To address this pressing need, Cisco has launched the AI Defense Explorer Edition, a self-service solution that enables organizations to assess and test their AI models against potential threats before deployment. This capability empowers developers and security teams to ensure that AI agents can withstand various attacks, such as prompt injection or jailbreak attempts. Key features of this edition include:

• - Dynamic Adversarial Testing: Conduct extensive testing of models and applications relevant to agentic workflows.
• - Robust Security Reporting: Obtain actionable insights that can be exported for compliance and auditing purposes.
• - Collaboration Tools: Facilitate teamwork among developers, enhancing the overall strength of security measures.

Additionally, Cisco is rolling out the Agent Runtime SDK, which allows organizations to integrate policy enforcement directly into agent workflows during development. This ensures that agents are secure from the get-go, aligning with various frameworks, including AWS and Google.

AI Security Integration in the SOC

Security Operations Centers (SOCs) are evolving due to the introduction of specialized AI agents designed to automate workflows, reducing the time spent on manual investigations. Instead of simply surfacing data, these agents actively evaluate and execute cybersecurity measures. Innovations from Cisco's partner Splunk enhance SOC capabilities by providing:

• - Exposure Analytics: An up-to-date inventory of assets, real-time risk insights, and visibility across the board.
• - Detection Studio: A streamlined workspace for the detection engineering lifecycle, facilitating the efficient management of threat detection.
• - Federated Search: A comprehensive search tool for correlating data across diverse environments, which is essential for quick and accurate threat assessments.

As Ryan Morris, President of Blackwood, highlighted, transforming the SOC from reactive to proactive is crucial in today's dynamic threat landscape.

Conclusion: A Secure Future for AI Deployment

Cisco's commitment to building a secure environment for AI agents represents a significant step in ensuring organizations can embrace AI without compromising security. The proactive measures, advanced tools, and adherence to Zero Trust principles create a solid foundation for the future of AI deployment in workplaces. With these innovations, Cisco showcases its dedication to aiding enterprises in navigating the complexities of the AI economy, enabling them to harness the incredible potential of AI safely.

For more details on Cisco’s latest security solutions for AI, visit their official website.