Netskope has launched AI Command Centre, extending its Netskope One AI Security suite.
The tool is intended to give security teams a clearer view of artificial intelligence use across corporate environments, including approved and unapproved services, on-premises systems and cloud deployments. Netskope also said it is joining Anthropic's Project Glasswing, which gives it access to Mythos.
Netskope is entering a market shaped by rapid enterprise adoption of AI tools and growing concern about governance. Research cited by the company found that only 6% of security teams reported full visibility over all AI deployments and usage within their organisation, while 94% said they had gaps in visibility into AI activity.
Visibility gaps
Data from Netskope Threat Labs pointed to a sharp rise in the use of AI software inside large organisations. The average enterprise it tracks recorded a fivefold increase in AI applications over the past year, tripled its AI user base, managed 37 deployed AI agents and saw 223 AI data policy violations each month, according to Netskope.
AI Command Centre is designed to discover AI assets across an organisation and connect that information to existing security context, including user identities, data stores and application trust ratings. Netskope said the system then highlights risk and suggests actions, including policy changes, remediation workflows or investigations, from the same interface.
Alongside the core release, Netskope outlined two additional discovery functions. One extends the Netskope One Client on managed endpoints to scan installed applications, active processes and listening ports to identify known AI agents, local models and browser extensions.
The second is a server discovery component built around a lightweight eBPF agent. Netskope said it intercepts TLS-encrypted AI traffic at kernel level on corporate virtual machines and Kubernetes nodes to identify AI infrastructure operating within the corporate perimeter.
Discovery tools
The launch also includes a new AgentSkope AI Risk AISecOps agent integrated with AI Command Centre. Netskope described it as an autonomous layer for triage, investigation and response, intended to help security teams handle AI-related incidents without adding staff.
AI Command Centre is generally available. Enhanced functions including endpoint AI discovery, server AI discovery, AI asset mapping and risk correlation, and the AI Risk AISecOps Agent are moving from private preview to general availability during the third quarter, Netskope said.
The announcement reflects a wider shift among cybersecurity suppliers to build products focused on AI oversight as companies deploy chatbots, coding assistants, autonomous agents and local models at speed. One of the central problems for security teams is that AI use often spreads beyond formal procurement channels, making it difficult to know which systems are handling company data and how they connect to existing infrastructure.
Growing risks
That challenge is particularly acute when employees bring in consumer AI tools or install browser-based assistants without approval. In those cases, defenders may have limited visibility into what data is being shared, whether models are hosted internally or externally, and how AI services interact with corporate systems.
Netskope argued that linking AI discovery to identity, application and data context is necessary for security teams to distinguish between low-level noise and issues that need action. It is positioning the new product as a way to consolidate that visibility rather than relying on separate tools for network traffic, endpoints and cloud applications.
For Netskope, joining Anthropic's Project Glasswing adds another strand to its AI security strategy. The company said participation gives it access to Mythos, although the announcement did not provide further technical detail.
Industry view
Sanjay Beri, set out the company's view of the problem the product is meant to address.
"Organisations have adopted AI faster than any security team can manually track, triage, or contain, and the tools nobody approved are almost always the ones carrying the highest risk. Netskope One AI Command Centre gives security teams the unified operational view they have never had before: every AI asset in the environment, how it's connected, what data it's touching, and what to do about it. They can now say yes to team members who want to take full advantage of AI, and do it securely. Paired with our new AI Risk AISecOps Agent, which reasons across the full context of every incident and closes the gap between knowing and doing, we're delivering a fundamental shift from security teams that react to AI risk, to security operations that anticipate and eliminate it," said Beri.
Jennifer Glenn pointed to the operational strain facing many organisations.
"Enterprise AI adoption has skyrocketed. Data volume and sprawl have created a pervasive visibility gap for security teams. For many organisations effectively correlating risk across managed and shadow AI assets, user identities and data stores is difficult. Addressing this challenge requires moving beyond siloed tools to a unified intelligence layer. Platforms that combine comprehensive AI discovery with real-time risk correlation are essential for enabling security operations to anticipate, prioritise, and autonomously eliminate AI-fueled threats at the speed the landscape demands," said Glenn.