RSA Secures AI Agents, MCP Servers with New Agent ID Platform

Key Takeaways

• RSA has introduced RSA Agent ID, a new identity security platform designed to manage and secure AI agents within regulated organizations.
• The platform addresses the growing challenge of “shadow AI” by providing discovery, security, and governance capabilities for AI agents and Model Context Protocol (MCP) servers.
• It aims to prevent data loss, operational disruption, and regulatory penalties by treating AI agents as distinct identities requiring robust access controls and lifecycle management.
• RSA Agent ID will be released in modules, with Discover and Secure available in November 2026, and Govern following in the first half of 2027.

RSA Unveils Agent ID Platform to Secure Enterprise AI Agents

RSA has announced the launch of RSA Agent ID, an innovative identity security platform engineered to help regulated organizations identify, secure, and manage their artificial intelligence (AI) agents. This new offering targets sectors such as financial services, government entities, and critical infrastructure operators, where stringent controls over AI agent access and actions are paramount.

The platform directly confronts a significant challenge emerging with widespread enterprise AI adoption: AI agents frequently operate with credentials and access to internal systems, performing automated tasks without the rigorous registration, ownership, and access review processes typically applied to human users. RSA emphasizes that AI agents, by virtue of holding credentials and entitlements, should be treated as distinct identities. Organizations often struggle with the sheer scale of identifying, assigning ownership to, and revoking access for every deployed agent.

This issue is projected to escalate dramatically. Gartner anticipates that by 2028, a typical Global Fortune 500 enterprise could be managing approximately 150,000 AI agents, a stark increase from fewer than 15 in 2025. Compounding this, research cited by RSA indicates that only 13% of organizations feel adequately prepared with the necessary governance for AI agents. This gap contributes to “shadow AI”—AI tools and agents operating outside approved security, compliance, and IT frameworks—a major concern for organizational security.

Addressing the Risks of Shadow AI

RSA’s solution brief states that incidents stemming from shadow AI can lead to severe consequences, including data breaches, operational disruptions, significant regulatory fines, and elevated breach recovery costs. The average cost of a security incident involving shadow AI is estimated at $5.39 million, exceeding the average data breach cost by approximately $400,000.

For highly regulated sectors like banking, public administration, and critical infrastructure, an unmanaged AI agent poses substantial risks. Such an agent could potentially access sensitive customer data, initiate financial transactions, exploit privileged cloud credentials, or interact with classified government systems. RSA asserts that these environments demand more robust controls than a cloud-only model can typically provide.

RSA Agent ID is available either as a unified platform or through three distinct modules:

The RSA Agent ID Discover module is engineered to locate AI agents and Model Context Protocol (MCP) servers across diverse data sources, including identity systems, cloud environments, endpoints, and gateways. It establishes a comprehensive registry for both known and previously unknown agents, assigning a human owner to each and recording its risk tier and lifecycle status.

RSA Agent ID Secure module applies policy enforcement to agent calls channeled through an AI/MCP Gateway. This gateway can be deployed in an RSA-hosted environment or within a customer’s own cloud, hybrid, or on-premises infrastructure. Organizations gain the ability to mandate authenticated human approval for high-risk agent actions, such as initiating wire transfers, accessing sensitive accounts, interacting with classified data, or handling personally identifiable information.

The RSA Agent ID Govern module focuses on comprehensive lifecycle controls. It facilitates continuous certification, risk-based access reviews, granular entitlement controls, and automated decommissioning processes. This ensures that agents do not retain access privileges beyond their designated tasks or business roles.

RSA offers sovereign control, allowing customers to dictate the deployment location of the gateway and where policy decisions are made. This ensures that enforcement actions and the generation of evidence remain within customer-controlled environments. RSA released a platform that generates tamper-evident records of agent actions, access decisions, and gateway calls. This evidence can be streamed to SIEM platforms and is mapped against 10 compliance frameworks, including NIST AI RMF 1.0, ISO/IEC 42001, DORA, and NYDFS Part 500.

RSA Agent ID Discover and Secure are slated for general availability on November 16, 2026. The RSA Agent ID Govern module is expected to become generally available during the first half of 2027.

What You Should Do

• Inventory AI Agents: Begin identifying all AI agents operating within your organization, including those in shadow IT environments. Document their purpose, access privileges, and data interactions.
• Establish Ownership: Assign a clear human owner to each AI agent, responsible for its lifecycle, security, and compliance.
• Implement Access Controls: Apply robust identity and access management (IAM) principles to AI agents, similar to human users. Grant least privilege and regularly review entitlements.
• Define Policy for High-Risk Actions: Develop and enforce policies requiring human oversight and approval for AI agent actions deemed high-risk or sensitive.
• Plan for Lifecycle Management: Establish processes for the secure provisioning, ongoing management, and timely decommissioning of AI agents, ensuring access is revoked when no longer needed.
• Monitor and Audit: Continuously monitor AI agent activities for anomalous behavior and ensure audit trails are captured and integrated into your SIEM for compliance and incident response.

Disclaimer: HackersRadar reports on cybersecurity threats and incidents for informational and awareness purposes only. We do not engage in hacking activities, data exfiltration, or the hosting or distribution of stolen or leaked information. All content is based on publicly available sources.