SentinelOne, an AI-native cybersecurity company, announced its vision and roadmap for AI security at OneCon 2025. The company revealed a new portfolio of solutions to secure AI systems, AI tools, and autonomous agents, while automating security operations.

SentinelOne reported that its AI Security strategy aims to help organizations accelerate and derisk their AI initiatives. The announcements at OneCon 2025 focused on securing AI systems today while preparing for autonomous security in the future.

The company introduced a new portfolio for securing AI, including:

• Prompt Security for Employees – Provides real-time visibility and control over employee GenAI usage, detecting shadow AI risks and preventing sensitive data exposure.
• Prompt Security for AI Code Assistants – Instantly redacts secrets, PII, and IP from AI-generated code, and uses a Vulnerable Code Scanner to block insecure outputs before production.
• Prompt Security for AI Applications – Protects custom-built AI solutions, from chatbots to complex automation, against emerging threats.
• Prompt Security for Agentic AI (Beta) – Offers real-time visibility, risk assessment, and governance for autonomous AI agents built on the Model Context Protocol (MCP).

In addition, following its acquisition of Observo AI, SentinelOne revealed an integration with Singularity AI SIEM. This AI-ready data pipeline combines intelligent streaming data control with agentic AI analytics and orchestration. The integration allows organizations to optimize data collection, enhance threat detection, and automate responses across security operations.

SentinelOne also expanded capabilities for Purple AI, its agentic security analyst. New features include in-line agentic auto-investigations with dynamic reasoning, automated investigations through Singularity Hyperautomation, and agentic custom detection rule creation. These updates aim to accelerate detection, investigation, and response from hours to minutes while enabling analysts to focus on high-priority incidents.

The company reported that Purple AI can now extend to third-party and custom AI applications via the Purple AI MCP Server. This provides secure integration between the Singularity Platform and any AI framework or large language model. SentinelOne noted that the open-source MCP Server is available today on GitHub.

Tomer Weingarten, CEO of SentinelOne, said, “Enterprises should capitalize on AI with confidence. That means mastering both Security for AI and AI for Security. At OneCon 2025, we are delivering the tools, strategies, and services organizations need to build, secure, and leverage AI safely today.”

Related post

View all