SailPoint Navigate Focuses on Identity Security Risks From Enterprise AI Agents

As enterprise organizations deploy autonomous AI agents across internal systems, security teams face growing pressure to adapt traditional access control frameworks. Identity management vendor SailPoint is centering its Navigate conference, taking place Oct. 5–8 in Austin, Texas, around this transition. Under the event theme "AI, secured," the gathering addresses how security architectures must evolve as task-oriented AI software joins human workers and traditional machine identities across enterprise environments, according to reporting by SiliconANGLE (https://siliconangle.com/2026/10/04/ai-agents-identity-security-sailpointnavigate/).

Unlike conventional identity governance models built around static roles and scheduled audits, agentic AI operates dynamically across systems and databases at machine speeds. As Krista Case, principal analyst and cyber resilience practice lead at theCUBE Research, observed, autonomous agents turn identity governance into an active execution challenge. Security platforms must maintain context around who or what is acting, whose authority the agent carries, and whether permissions should persist as tasks evolve across the agent's lifecycle.

Data indicates a substantial visibility deficit surrounding these non-human accounts. SailPoint research cited by ZK Research principal analyst Zeus Kerravala found that 97% of AI agents have access to sensitive corporate data, while only 21% of organizations express high confidence in their ability to manage AI agent security risks. Kerravala also pointed to a SailPoint proof of concept at a Fortune 500 company that uncovered more than 10,000 previously unknown AI agents operating within its infrastructure.

To mitigate these blind spots, identity vendors are shifting away from periodic compliance checks toward active risk remediation. Chandra Gnanasambandam, SailPoint's executive vice president of product and chief technology officer, noted in August that the company is moving toward a continuous security loop that combines discovery, lifecycle policy governance, and real-time remediation to neutralize attack vectors before exploitation.

Managing agentic AI is as much an organizational hurdle as a technical one. Case noted that software agents can be spun up faster than traditional access management workflows can assign ownership and establish boundaries. As a result, enterprises need operating models that integrate AI developers, cybersecurity teams, identity managers, and business unit leaders before agent deployments outpace governance capacity.

Analysts also emphasize that while runtime containment tools such as sandboxes restrict operational boundaries, they do not replace identity governance. Kerravala pointed out that identity serves as the foundational system of record across infrastructure, identifying agent creators, assigning accountable human owners, and tracking permission lifetimes.

SailPoint is positioning identity as a primary control plane for managing combined human, machine, and software entities. SailPoint founder and chief executive officer Mark McClain highlighted that proliferation of AI agents introduces a new class of non-human identities, expanding the enterprise attack surface. To address this, SailPoint is collaborating with Amazon Web Services on a unified identity plane. However, Case cautioned against relying on any single vendor platform as a sole control layer, noting that effective agent governance requires coordination across identity platforms, runtime infrastructure, applications, and security telemetry.