Netskope has announced a new set of security capabilities for AI called Netskope One AI Security. The functionality is designed to give organizations greater insight and control over the use of AI applications, models, and agents within their IT environment.
According to Netskope, the security suite is integrated into the Netskope One platform and designed to protect various components of the AI ecosystem. These include AI applications, AI agents, datasets, and users in both public SaaS environments and private or internally hosted AI systems. Workflows in which autonomous AI agents communicate with other systems are also covered by the security.
In addition to the security capabilities, Netskope is also introducing an AI Index. This index is intended to provide global insight into the extent to which organizations are applying AI and the risks that this entails.
The announcement includes four new products that focus on different aspects of AI security. Netskope One Agentic Broker is designed to provide visibility and control over communications between AI agents and data sources, including transactions outside of approved applications. This allows organizations to better monitor how autonomous AI systems access company data.
With Netskope One AI Guardrails, the company aims to protect organizations against AI-specific attacks, such as prompt injection and attempts to circumvent language model security mechanisms. The functionality acts as a moderation system for interactions with AI models and is designed to detect and block risky or undesirable content.
In addition, Netskope is introducing One AI Gateway for organizations that host their own AI applications or language models. The gateway inspects traffic between applications and AI models and enforces security policies in private AI environments, such as on-premises or in a virtual private cloud.
The fourth component, Netskope One AI Red Teaming, focuses on testing AI systems during the development phase. AI models are exposed to simulated attacks to identify vulnerabilities and security risks before systems go into production.
Growing AI investments and new risks
The introduction of the security suite comes at a time when organizations are increasingly investing in AI. According to research firm IDC, global business spending on AI will reach approximately $241.8 billion in 2025. That amount is expected to rise to more than $867 billion in 2029.
This rapid adoption brings new security challenges. Organizations that still use traditional security solutions often have limited visibility into how data is used within AI applications, models, and agents. As a result, AI systems can operate outside existing security controls and introduce risks such as data breaches, manipulated prompts, or unwanted use of company data.
Platform for visibility and control
With Netskope One AI Security, the company aims to offer organizations a centralized way to map and manage AI usage. The platform is designed to detect AI applications, models, and agents, analyze risks, and enforce security measures in real time.
The suite works in conjunction with Netskope NewEdge AI Fast Path, an infrastructure layer designed to optimize network performance for AI services. By combining this infrastructure with zero trust security controls, Netskope aims to improve both the performance and security of AI workloads.
According to co-founder and CEO Sanjay Beri, the rapid growth of AI requires new network and security architectures. He believes that the combination of the Netskope One platform and the NewEdge network should form an infrastructure specifically designed for AI-driven environments and the complex data flows that come with them.
Customers also emphasize the importance of governance around AI. Alexander Schuchman, CISO at Colgate-Palmolive, says that AI can help organizations improve processes and accelerate innovation, but that its use requires clear governance and ethical guidelines.
The new products Netskope One Agentic Broker, Netskope One AI Gateway, Netskope One AI Guardrails, and Netskope One AI Red Teaming are available immediately, according to the company. Netskope will demonstrate the Netskope One platform later this month at the RSA Conference in San Francisco.