Startups & Funding

Vero Security Raises $11 Million Seed to Enforce AI Agent Policies at Runtime

Vero Security raised an $11 million seed led by TLV Partners to enforce AI security policies at runtime, translating one Fortune 1000 customer's rule into 15+ live controls in minutes.

Vero Security Raises $11 Million Seed Round to Enforce Policies Across AI Agents
Vero Security Raises $11 Million Seed Round to Enforce Policies Across AI AgentsPexels / pixabay
By James Calloway6 min read

Why it matters

  • Vero Security raised an $11 million seed round led by TLV Partners, with GTM Capital, Maple Capital and angels participating.
  • Founded in 2025, Vero already claims multiple Fortune 500 customers and has offices in New York and Tel Aviv.
  • One Fortune 1000 customer's single policy was translated into more than 15 controls within minutes, per Vero.
  • Founders Elad Tessler and Daniel Meirov met in Unit 8200, co-leading a cyber operations unit of over 100 people.
  • Competitors such as Sysdig's AI Defense also target runtime observation and enforcement of AI agent behavior.

Vero Security has raised an $11 million seed round to turn enterprise AI security policies into controls that operate while employees' AI tools and agents are actually working. TLV Partners led the round, with participation from GTM Capital, Maple Capital, and angel investors. Founded in 2025, the company already claims multiple Fortune 500 customers.

The bet behind the funding is specific: runtime enforcement of AI agent behavior becomes a permanent operational function, not a one-time assessment. Vero positions itself in a market it calls AI Security Engineering — the continuous work of discovering how AI is being used, assessing risks, and keeping protections current as tools and permissions change.

The stakes are straightforward. AI agents that can read company files, connect to repositories, invoke tools, and execute code need boundaries that stay effective as those capabilities evolve. Vero plans to spend the new capital on product development, broader coverage of enterprise AI technologies and security workflows, and expansion of its U.S. commercial presence.

What does Vero actually enforce?

Vero's pitch targets a common complaint from security teams: assessment tools that produce findings lists without implementing fixes. The company says its platform takes on more of the operational workload instead.

The workflow runs as follows, according to the company:

  • Security professionals define the policy and decide which changes the platform may implement automatically.
  • Vero translates approved decisions into controls and coordinates fixes with the relevant owners.
  • The platform checks that the work has been completed.
  • Protections adjust as tools and permissions evolve; changes requiring a fresh risk decision return to a person, with supporting analysis and a recommended action.

The scale of the engineering hidden inside a single policy statement is visible in one customer example. Vero says a Fortune 1000 customer started with a simple rule — only developers may connect agents to corporate code repositories or execute code. The platform translated that one statement into more than 15 controls spanning coding agents, identities, Model Context Protocol servers, skills, and runtime actions, within minutes.

MCP, in short, is a connection mechanism that lets AI applications interact with external tools and data sources; skills supply reusable instructions or procedures. Together with an agent's credentials, these components determine what the agent can actually do. Enforcing who may use an agent is only part of the task — the connected tools, credentials, and operations must also reflect that decision.

What is the AI Runtime Governor?

At the center of the platform sits Vero's AI Runtime Governor. The company says it inspects supported prompts, tool calls, and agent actions, applying policy when an operation is attempted rather than waiting for a later manual review.

The listed interventions include:

  • Preventing a coding agent from reading unrelated secrets.
  • Blocking actions that fall outside an approved policy.
  • Stopping malicious skills or prompt injection from triggering unauthorized behavior.

Prompt injection occurs when material an AI system encounters contains instructions intended to redirect its behavior. The consequences scale with the agent's reach: for an assistant that only drafts text, the result may be an unreliable answer. An agent with repository access or permission to execute commands can turn the same manipulation into an operation against a real system.

This distinction matters for how enforcement actually works. A rule written into an agent's instructions asks the software to avoid a sensitive resource — it expresses intended behavior. An enforcement mechanism, by contrast, must inspect the attempted access and apply the boundary along the execution path it supports. The gap between those two approaches is the gap Vero is selling against.

Why configuration checks alone fall short

Vero's website describes four connected functions: inventory, posture analysis, behavior detection, and enforcement. Inventory identifies agents and their owners; posture analysis examines configuration and reachable credentials; behavior detection examines activity during a session; enforcement applies the policy.

The company's illustrated workflow follows an agent whose configuration exposes production credentials and whose approved skill later receives a malicious update. Vero describes addressing the configuration problem first, then blocking the harmful call when the changed skill tries to act.

The underlying logic is that an initial approval is only a snapshot. A tool update, a new connection, or an additional permission can change the risk profile after deployment. Continuous monitoring and action controls address a different moment from the original assessment — which is precisely why Vero argues that point-in-time configuration audits cannot carry the load alone.

There is a caveat the company itself acknowledges in its framing. Vero's enforcement claims apply to supported prompts, tools, and actions, meaning the practical scope of any deployment follows the integrations and execution paths the platform covers. Enterprises evaluating it will need to ask what their agents actually run on, not just what the platform supports in principle.

On deployment friction, Vero says it operates through the endpoint, identity, AI, and security tools an enterprise already uses, without requiring employees to install anything new. That claim reduces adoption barriers; how complete the coverage is across heterogeneous enterprise stacks remains the open question.

Who is behind Vero, and why did investors commit?

Vero was founded by Elad Tessler, its CEO, and Daniel Meirov, its CTO. Tessler previously led product at Astrix Security. Meirov was Sightfull's first employee and led its R&D and product work. The founders met in Unit 8200, Israel's signals intelligence unit, where they co-led a cyber operations unit of more than 100 people, according to the announcement. Vero has offices in New York and Tel Aviv.

GTM Capital says it invested after a structured evaluation involving its community of Fortune 500 CIOs and CISOs. The announcement describes those executives working directly with Vero before the firm committed — an unusual diligence path that suggests demand validation from target buyers rather than purely financial conviction.

How crowded is this market?

Vero is not operating alone. Runtime controls are becoming a more prominent part of enterprise AI security, and other launches — including Sysdig's AI Defense — also focus on observing agent behavior and enforcing boundaries during execution. Vero's funding pitch differentiates on the ongoing operating work around those boundaries: implementing decisions, verifying remediation, and keeping controls aligned with policy over time.

Whether that operational framing constitutes a durable moat is unproven at seed stage. A Fortune 1000 customer example and unnamed Fortune 500 users are adoption signals, but the company has not disclosed revenue, retention, or how many execution paths the Runtime Governor actually covers in production.

Alongside the funding, Vero is launching AISecurityEngineering.org, an open community and knowledge center with field guides, workshops, and an ASE credential track. The initiative targets professionals who are establishing AI security practices while often carrying other responsibilities inside their organizations — a common staffing reality in enterprises that have not yet built dedicated AI security teams.

The broader opportunity, as the company frames it, is making security policy durable even when the software underneath it keeps changing. As enterprises give agents more useful access, the test for platforms such as Vero is whether they can consistently carry a security team's decisions through to the actions those agents attempt.

Source: unite.ai

Share this article:

More from James Calloway

James Calloway

Show full bio

News editor covering people, launches and day-to-day practice in artificial intelligence and machine learning at AI In Context.

722 articles

Related articles

  1. Vanta Signs Deal to Buy AI Security Firm Pillar Security
  2. OpenAI to Acquire Promptfoo and Build Red-Teaming Into Frontier
  3. Enterprise AI Guardrails Are Shifting From Frontier Fear to Buyer Contracts
  4. OpenAI Launches Initiative for Democratic AI Oversight in National Security
  5. Vijil CEO Vin Sharma: AI Agents Need Trust Infrastructure Before Autonomy

« Previous articleNext article »