Qevlar Raises $14M to Lead the Agentic AI Revolution
As cyberthreats escalate in frequency, sophistication and impact, security operations centers (SOCs) are being pushed to their breaking point. Human analysts, already in short supply, are drowning in alerts (most of which are false positives) and rely on outdated, manual investigative workflows that can’t keep up with the pace of today’s attacks.
Enter Qevlar AI, a Paris-based cybersecurity startup charting a bold new path with its autonomous, agentic AI platform. This week, the company announced it has raised $14 million in total funding, including a fresh $10 million round led by EQT Ventures and Forgepoint Capital International. The round also attracted strategic angel backing from Olivier Pomel (CEO, Datadog), Florian Douetteau (CEO, Dataiku) and Mehdi Ghissassi (former director of product, Google DeepMind), all of whom see Qevlar’s vision as a transformative leap forward in modern security operations.
“Our mission is simple but urgent: To shift SOCs from reactive alert factories to proactive threat-hunting teams,” said Qevlar AI co-founder and CEO Ahmed Achchak. “We’re not just augmenting analysts; we’re solving the impossible math problems they face every day.”
The Impossible Equation of Cybersecurity
The vast majority of SOCs are overburdened and under-resourced, with analysts unable to respond to 67% of daily alerts, with an estimated 83% of those alerts ultimately proving to be false positives. The ongoing cybersecurity talent shortage only worsens the situation. The World Economic Forum reports that two-thirds of organizations face moderate-to-critical skills gaps, and just 14% say they have the necessary people and capabilities.
Qevlar AI’s answer is not another new dashboard or another SIEM plug-in. Instead, the company offers a fully autonomous investigation platform, powered by agentic AI, that mimics the way human analysts think but operates at scales and speeds that are orders of magnitude more efficient.
So, What is Agentic AI?
Simply put, agentic AI is like a super-smart assistant that doesn’t just wait for instructions. It identifies problems, makes decisions and takes actions on its own. Instead of offering suggestions like a chatbot or co-pilot, it does the work. In cybersecurity, that means investigating threats, gathering evidence and resolving routine cases without human input.
But that doesn’t mean humans are left out of the loop. Qevlar’s agentic AI streamlines the process in two critical ways. It dramatically improves mean time to detection (MTTD) by surfacing real threats faster, and it enables humans to engage sooner and more effectively, which ultimately leads to reducing mean time to resolution (MTTR) for high-risk, high-impact cases that truly require expert judgment.
“Our agents don’t suggest actions, they take them,” said Achchak. “They investigate alerts, gather evidence, make contextual decisions and even close tickets when appropriate. In less than three minutes, they do what takes a human analyst 30 to 60 minutes.”
From Human Limits to Machine Precision
Since its launch in 2023, Qevlar AI has already delivered measurable impact across several enterprise customers and managed security service providers (MSSPs). The results are compelling:
- Alert investigation time reduced from 40 minutes to just three minutes
- L1 and L2 analyst workloads reduced by 90%
- Alert classification accuracy improved to 99.8%, surpassing the 97% ceiling typically achieved by human experts
- 100% of non-harmful alerts autonomously closed without human review
The platform is built on a proprietary knowledge graph and large language model (LLM) stack designed specifically for security operations. These components enable each autonomous agent to execute full-loop investigations while maintaining traceability and auditability, an essential requirement in today’s high-stakes cybersecurity environments.
Born of Machine Learning, Not Retrofitted Into It
While most cybersecurity startups (and even large, well-established tech and security companies) are scrambling to bolt AI onto existing workflows, Qevlar stands apart by approaching the problem from the opposite direction.
“We come at this from the machine learning side of the equation. That’s our DNA, even before the GenAI hype cycle,” said Achchak. “Most players in this space are coming from cybersecurity backgrounds and trying to learn ML or LLMs. For us, it’s the opposite. We’ve been deep in this for years, and now we’re applying it to solve one of the most painful security problems there is.”
Qevlar’s ML-first mindset results in not just speed and autonomy, but also safety. “Because we come from the machine learning world, we’ve always approached this with a deep understanding of model behavior, edge cases and how things can go wrong,” he said. “That means when we work with LLMs, it’s not just bolt-on GenAI, it’s secure by design, and in turn, we are a very important element of secure software supply chains (SSSC). We anticipate failure modes, we structure the context, and we control the outputs in ways others don’t even think about.”
This orientation gives Qevlar a head start, not just in technology, but in culture and capability. Its team isn’t adapting to AI, it was born of it. The result is an investigation agent that behaves more like a seasoned, tireless analyst than a glorified chatbot.
“Our goal isn’t to build co-pilots,” he added. “It’s to create autonomous agents that can think, act and deliver outcomes at speed and at scale.”
Investors and Customers Aligned
EQT Ventures and Forgepoint Capital International see Qevlar AI as a foundational building block in the future of cybersecurity. Damien Henault, managing director at Forgepoint, will join Qevlar’s board of directors as part of the funding round.
“Qevlar AI’s autonomous agent-led investigations are reshaping the future of SOCs,” said Henault in a statement. “By empowering L1 and L2 analysts to focus on higher-value work, they’re helping solve one of cybersecurity’s most critical pain points.”
Real-world deployments affirm this vision. Nomios, a major European MSSP with over 60 SOC analysts, has adopted Qevlar AI to cut investigation times and improve analysis consistency. “With Qevlar AI, we can rapidly analyze even the most complex of cases in just three minutes,” said Nomios COO Eric Bohec.
Similarly, Daniel Aldstam, chief information security officer, head of group security of Nordic digital infrastructure provider GlobalConnect, praised the platform’s seamless integration and operational impact. “Adding something new is always a risk, but Qevlar AI plugged into our environment effortlessly and started delivering value right away.”
Scaling the Future of Security Operations
The new investment will be used to expand the Qevlar AI team, which has already quadrupled in the past year, and to drive further research and development. International expansion is also a key priority, especially in North America, where SOC workloads are among the heaviest.
Qevlar AI will showcase its platform at the RSAC 2025 Conference in San Francisco from April 28 to May 1, appearing in the Early-Stage Expo (booth ESE 16). For attendees, it will be a chance to see firsthand how Qevlar’s agentic AI is transforming one of the most human-intensive fields in enterprise technology.
As for what’s next, Ahmed was clear: “We’re just getting started. The future of security operations is autonomous and Qevlar intends to lead that future.”
