19 Nov 2026, Thu

VeroSecurity Emerges from Stealth with $11 Million Seed Round to Define AI Security Engineering

VeroSecurity, a startup founded in 2025 by IDF 8200 alumni Elad Tessler and Daniel Meirov, has officially emerged from stealth mode, announcing an $11 million seed funding round led by TLV Partners. The company is setting its sights on a critical, yet largely uncharted frontier in the technology landscape: AI Security Engineering (ASE). As organizations across the globe scramble to integrate generative AI tools and autonomous agents into their internal workflows, the traditional cybersecurity playbook is proving insufficient. VeroSecurity intends to use this new capital infusion to accelerate the development of its specialized platform and scale its go-to-market operations, particularly within the United States.

The Rise of AI Security Engineering

The term "AI Security Engineering" represents more than just a buzzword; it is an acknowledgment that the rapid adoption of AI tools by employees—often without the oversight of centralized IT departments—has created a new, distinct category of risk. Unlike traditional cybersecurity, which focuses on protecting networks, endpoints, and cloud infrastructure from external threats, ASE is concerned with the governance, safety, and operational integrity of AI models and the agents that execute tasks on behalf of a company.

The challenge is a moving target. Because employees are frequently experimenting with, modifying, and chaining together various AI tools, the enterprise attack surface is expanding and shifting in real-time. Standard security tools, which rely on static rules or historical threat patterns, cannot keep pace with the high-velocity, automated nature of AI agents. Recognizing this gap, VeroSecurity is positioning itself not merely as a vendor, but as a catalyst for the formalization of this new discipline.

Moving Beyond the Dashboard

At the heart of the VeroSecurity philosophy is a rejection of the "dashboard-first" approach that has defined the previous generation of security tools. Elad Tessler, the company’s CEO, argues that CISOs and their teams are suffering from information overload. Current platforms are often designed to aggregate massive amounts of data, presenting them in complex tables and charts that tell security professionals what is wrong without providing the tools to actually fix it. This leaves the burden of remediation entirely on the shoulders of already stretched security teams.

"CISOs have seen enough dashboards and tables that tell their teams what’s wrong and leave the work to them," says Tessler. "AI Security Engineering deserves a better operating model, designed for AI rather than borrowed from other areas of security."

VeroSecurity aims to break this cycle by shifting the focus from monitoring to active, automated enforcement. The company is working closely with the first generation of practitioners who are currently grappling with the challenges of securing enterprise AI, ensuring that the platform is built to solve the practical, day-to-day problems these professionals face. The goal is to create a system that handles the heavy lifting of security operations—keeping protections current as AI tools evolve—while reserving human attention only for high-stakes decisions that require nuanced judgment.

The AI Runtime Governor

The technical centerpiece of the Vero platform is the "AI Runtime Governor." This engine operates by embedding security decisions directly into the operational flow of AI tools. By enforcing pre-approved policies in real-time, the Governor can intercept and block unauthorized agent behaviors before they result in a data breach or operational disruption.

This is a critical distinction in the era of AI. Because agents operate at machine speed, reviewing their actions after the fact—as is standard with traditional log-based auditing—is often too late to prevent damage. The AI Runtime Governor provides a proactive layer of defense, ensuring that as agents interact with sensitive data or external APIs, they remain within the guardrails established by the security team.

When a situation arises that falls outside of existing policy or requires a risk assessment that the system cannot autonomously resolve, the platform presents the relevant data, context, and a recommended course of action to the human security engineer. This allows the human to focus on policy-making and strategic risk management rather than the endless, repetitive task of monitoring individual agent actions.

Establishing a New Discipline

Beyond the product itself, VeroSecurity is taking a proactive role in building the community around ASE. The company is launching AISecurityEngineering.org, an open community and knowledge hub designed to serve as a central repository for the discipline. This site will feature practical field guides, technical workshops, and a structured credential track, aiming to provide the training and body of knowledge that are currently missing from the industry.

The urgency for this, according to Tessler, is evidenced by the rapid growth in labor market demand. VeroSecurity has been tracking the rise of job postings specifically for "AI Security Engineers." Their data shows that the number of open roles with that specific title grew from 367 on October 25 to 451 on November 14—a 23% increase in just three weeks.

"Demand is arriving before the discipline has been written down," Tessler noted. He draws parallels to established security domains like Application Security (AppSec) and detection engineering, both of which spent years building a shared language and set of best practices before they became standard career paths. He believes AI Security Engineering is at a similar inflection point and that the industry needs to move quickly to define these standards to ensure the long-term security of enterprise AI deployments.

Solving the "Impossible" Workload

For a large enterprise, the volume of change in the AI landscape is overwhelming. Companies are frequently updating their agents, swapping out models, and integrating new tools on a weekly—or even daily—basis. Each change has the potential to impact the security posture of the entire organization. Managing this manually is simply not sustainable.

"The Vero platform undertakes the impossible part of the ASE work, leaving the human engineer in situ to make judgement calls where necessary," Tessler explained. The "impossible" aspect, he clarifies, is the sheer tempo and scale of change. Agents act in seconds, and they do so with a complexity that makes manual oversight impossible. The platform handles the mechanics of mapping policies to these tools, ensuring that security remains continuous and automated.

The human role, conversely, is elevated. Security engineers are left to determine what the business will accept in terms of risk, which exceptions are valid for specific business cases, and whether a new, high-potential AI tool is worth the security trade-offs.

"We’re betting on more AI security engineers, not fewer," said Tessler. "Which is why we’re building AISecurityEngineering.org to train them. Without Vero, keeping up with AI in a large company takes a team. With it, one person, or a security leader taking it on alongside their job, can stay ahead."

By providing both the automation platform and the educational foundation for the discipline, VeroSecurity is positioning itself to be a primary architect of how enterprises will secure their AI agents moving forward. The company’s focus on the intersection of automated runtime governance and the formalization of the AI Security Engineer role highlights a growing consensus that AI security is not merely a feature to be added to existing tools, but an entirely new engineering domain that requires a purpose-built approach. As enterprises continue to deepen their reliance on AI, the work being done at VeroSecurity to bring order to this evolving landscape will likely play a significant role in determining how successfully—and safely—these companies navigate the transition.

Leave a Reply

Your email address will not be published. Required fields are marked *