Human-Cyber-Physical Systems (HCPS) face an ever-growing number of cyber threats. Addressing these challenges effectively requires both time and specialized expertise, which are often in short supply.

RobustifAI propose automating Security Operations Center (SOC) processes using a human-centric approach powered by Large Language Models (LLMs). Our system handles four key activities:

• Threat Intelligence Synthesis
• Automated Intrusion Detection Rule Generation
• Incident Analysis and Reporting
• Response Plan Generation

Our system operates through a seamless workflow. Analysts input relevant threat information, which is then processed by a Retrieval-Augmented Generation (RAG) mechanism. This mechanism, driven by LLMs, retrieves and synthesizes qualified reports into reliable summaries. When new vulnerabilities, such as Common Vulnerability Exposure (CVE) bulletins, are identified, RobustifAI automatically generates updated intrusion detection rules.

The result is a cybersecurity framework that is faster, more reliable, and centered on human needs, ensuring robust protection for critical infrastructures in an ever-evolving threat landscape.