Putting Out the Fire and Finding What Started It: Qevlar AI’s Bet on Autonomous Security Intelligence

2026-03-17 · Source: The French Tech Journal · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cybersecurity & Data Privacy · Depth: Intermediate, medium

Summary

Qevlar AI, a Paris-based startup, recently secured \$30 million in funding, bringing its total to \$44 million, to advance its autonomous AI SOC platform beyond alert investigation into comprehensive security intelligence. The company addresses the "consistency problem" of large language models (LLMs) in cybersecurity by employing a hybrid approach of "expert systems," including Bayesian modeling and graph AI, to ensure reliable and explainable security decisions. This strategy has led to a claimed 10x reduction in investigation time to three minutes and 100% alert investigation for customers like Orange Cyberdefense and Sodexo, while offering predictable, per-investigation pricing. Qevlar AI's vision is to transition from reactive "firefighting" to proactive intelligence, correlating historical alerts to uncover root causes and patterns. Despite challenges in hiring, the company is rapidly expanding its European and U.S. presence with 52 employees.

Key takeaway

Qevlar AI's autonomous SOC platform addresses LLM inconsistency in cybersecurity by employing a hybrid AI approach combining Bayesian modeling, classical ML, and a Graph AI "dynamic ontology." This system achieves a 10x reduction in investigation time (3 minutes per alert) and 100% alert coverage, enabling reliable, explainable, and cost-predictable security intelligence. It shifts SOCs from reactive firefighting to proactive threat root cause analysis.

Topics

• AI in Cybersecurity
• Security Operations Center
• Large Language Models
• Graph AI
• Autonomous Security

Best for: AI Security Engineer, Security Engineer, Investor

Related on AIssential

• Continuous Offensive Security: The Line We've Been Walking — AI-driven offensive security is crucial for continuous defense against autonomous attackers and new AI-specific attack surfaces.
• Explainable AI-Driven Cyber Risk Analytics and Model Reliability Assessment for Intelligent Governance of U.S. Critical Infrastructure: An XGBoost and SHAP-Based Intrusion Detection Framework — An XAI-driven framework enhances cyber risk analytics and intrusion detection for critical infrastructure by improving model transparency and reliability.
• Exaforce raises $125M Series B to build AI for catching and stopping cyberattacks as they happen - TechCrunch — AI-driven security platforms automate threat detection and response, significantly reducing manual effort and false positives.
• AI Police Dog Security Simulation Earns a 23.64 Proof of Usefulness Score by Building an Autonomous Security Simulation for Industrial Zones — An autonomous security dog simulation uses FSM for reliable patrolling, pursuit, and energy-aware fail-safes in industrial zones.
• Anthropic's Mythos is evolving faster than expected, reports AI safety agency — AI model capabilities, particularly in cybersecurity, are advancing much faster than anticipated, even within minor version updates.
• IBM Announces New Cybersecurity Measures to Help Enterprises Confront Agentic Attacks — AI-powered offensive capabilities necessitate systemic, AI-powered defensive strategies operating at machine speed.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The French Tech Journal.