In cybersecurity, timing matters. The last half decade was defined by cloud-native security players like Wiz, which scaled by simplifying complex workflows and building elegant user experiences. But that playbook – layering incremental features on top of legacy approaches – is already showing signs of exhaustion.
Today, most founders and investors are still talking about AI as an “add-on.” A better detector here, a smarter alert there. That misses the real story. The next generation of cybersecurity companies won’t be about adding AI features. They’ll be agentic from the ground up: systems that monitor, decide, and act autonomously in real time.
And where better to lead this new frontier than from Israel?
Dashboards to agents
Talk to CISOs of global corporations, as we do regularly, and the frustration is consistent. They don’t want another dashboard or another set of alerts. They want systems that do something about it. They want tools that can identify an attack surface, simulate an exploit, and shut it down, not wait for a human analyst drowning in tickets.
That’s the promise of agentic security. Think of automated “co-pilots” not for developers but for defenders: agents that continuously evaluate vulnerabilities, adapt policies, and trigger remediation workflows.
What we’re seeing on the ground
Over the past 18 months, we’ve met dozens of founders pitching AI in security. Most are layering it on top of existing categories. A smaller set is attracting the strongest attention from both CISOs and investors by rebuilding the architecture entirely. They start with agents, workflows that run themselves, and data pipelines treated as part of the security boundary.
Early signs are already visible in Israel. Teams are designing contextual exposure management that prioritizes what really matters, or workflow automation frameworks that let micro-agents respond faster than human teams ever could. It’s a quiet shift, but it is happening now.
Prompt Security, the AI security start-up acquired by SentinelOne this past August, is a good example of this wave. Teams like Tonic Security, which focuses on contextual prioritization of exposures, and BlinkOps, which automates incident response through micro-agents, hint at how agentic security is already taking shape.
It matters and Israel will lead
Cybersecurity has always followed technology: networks, then endpoints, then cloud, then SaaS. AI is the next epoch, but it isn’t just another category to secure. It changes the very fabric of how security is delivered. In this new era, static dashboards are a liability. Autonomous agents are the only way to keep pace.
Globally, only a handful of start-ups are re-architecting security around agents. None have broken out at unicorn scale yet. Israel is uniquely positioned to get there first. The same ecosystem that produced global leaders in network, endpoint, and cloud security now fuses cyber DNA with world-class AI and DevOps talent. That combination is rare, and it’s exactly what agentic security demands.
The next cyber unicorn will not look like Wiz. It will not win by being simpler, friendlier, or faster to market alone. It will win by being agentic: by building autonomous systems that secure AI with AI. If history is any guide, it will be built here, in Israel.
The writer is a general partner at Hetz Ventures, where he leads the firm’s cybersecurity portfolio.