Netskope's AI Command Centre Tackles Enterprise AI Visibility Gaps

A striking 94% of security teams lack full visibility into AI deployments within their organizations, according to Netskope research. This visibility gap is particularly concerning given the rapid adoption of AI tools in enterprise environments, with the average organization experiencing a fivefold increase in AI applications and a tripled AI user base over the past year. The numbers are telling: 37 deployed AI agents and 223 monthly AI data policy violations per organization. It's clear that the lack of visibility into AI activity poses significant security risks, making it challenging for security teams to distinguish between low-level noise and issues that require immediate attention.

The AI Visibility Crisis in Enterprise Environments

The AI visibility crisis in enterprise environments is a pressing concern. The rapid proliferation of AI tools, both approved and unapproved, has created an environment where security teams struggle to keep pace. The use of AI software inside large organizations has skyrocketed, with a sharp rise in AI applications, AI user base expansion, and AI data policy violations. This has resulted in a pervasive visibility gap for security teams, making it difficult to correlate risk across managed and shadow AI assets, user identities, and data stores. In my analysis, this visibility gap is a significant vulnerability that can be exploited by malicious actors, highlighting the need for a comprehensive platform that links AI discovery to security operations.

Technical Breakdown: eBPF and TLS Interception for Server Discovery

From a technical perspective, Netskope's AI Command Centre utilizes eBPF kernel-level TLS interception for server discovery, providing a more comprehensive method for detecting AI infrastructure compared to traditional network monitoring methods. The eBPF agent intercepts TLS-encrypted AI traffic at the kernel level on corporate virtual machines and Kubernetes nodes, offering a deeper level of visibility into AI infrastructure operating within the corporate perimeter. This approach enables security teams to discover AI assets across the organization and connect that information to existing security context, including user identities, data stores, and application trust ratings. For instance, the eBPF agent can identify AI traffic that is encrypted or operating at the kernel level, providing security teams with a more accurate picture of their AI landscape.

The use of eBPF kernel-level TLS interception is a significant innovation in AI discovery, providing security teams with the visibility they need to manage AI risk effectively. By leveraging this technology, organizations can gain a better understanding of their AI infrastructure and take proactive steps to mitigate potential security threats.

Operational Shift: From Reactive to Proactive AI Risk Management

The launch of AI Command Centre marks a significant shift in the approach to AI risk management, from reactive to proactive. By providing a unified operational view of all AI assets in the environment, security teams can now anticipate and eliminate AI-fueled threats at the speed the landscape demands. The platform's ability to highlight risk and suggest actions, including policy changes, remediation workflows, or investigations, enables security teams to take a proactive approach to AI risk management. This is particularly important given the rapid evolution of AI threats, which can quickly exploit vulnerabilities in unapproved AI services. For example, the AI Risk AISecOps Agent can autonomously triage, investigate, and respond to AI-related incidents, reducing the burden on security teams and enabling them to focus on high-priority threats.

AI Loop Perspective: Why Unified AI Governance is Non-Negotiable Now

In my assessment, the lack of visibility into AI activity is a significant vulnerability that can be exploited by malicious actors. The use of AI software inside large organizations has created an environment where security teams struggle to keep pace, making it challenging to distinguish between low-level noise and issues that require immediate attention. The partnership between Netskope and Anthropic's Project Glasswing, which provides access to Mythos, is a significant development in this space. While the exact capabilities of Mythos are not specified, it can be inferred that it may enhance Netskope's AI visibility, possibly addressing gaps in AI governance and oversight. As the AI landscape continues to evolve, it's clear that unified AI governance is non-negotiable. Organizations must prioritize visibility into AI activity, leveraging platforms like AI Command Centre to manage AI risk effectively and ensure the secure adoption of AI tools.

The attack surface analysis reveals that the lack of visibility into AI activity is a significant vulnerability that can be exploited by malicious actors. As such, it's essential for organizations to prioritize unified AI governance, leveraging platforms like AI Command Centre to manage AI risk effectively. By providing a comprehensive view of AI assets and connecting that information to existing security context, AI Command Centre enables security teams to take a proactive approach to AI risk management. If this trend holds — and the data suggests it will — we are looking at a future where AI risk management is a top priority for security teams. The real innovation is not the use of AI itself, but the ability to manage AI risk effectively, and that changes everything.

---

— Alice Petrovna, Lead Cybersecurity Analyst & DevSecOps Expert at AI Loop