# Neo emerges with $100 million to police AI agents on enterprise endpoints

> Source: <https://runtimewire.com/article/neo-100-million-ai-agent-security>
> Published: 2026-07-20 17:18:24+00:00

[Neo](https://www.neo.ai/), founded by former SentinelOne executives Nick Warner and Shlomi Salem alongside EasySend co-founder Eran Shirazi, emerged from stealth on July 20th with $100 million in cumulative seed and Series A funding to control AI agents operating with employee permissions.

The Boston-based, American-Israeli cybersecurity company is backed by Andreessen Horowitz, Bessemer Venture Partners, Craft Ventures and Merlin Ventures. [Andreessen Horowitz said in a post on X](https://x.com/a16z/status/2079222562807005369) that it led Neo's seed round, arguing that endpoint detection, data-loss prevention and zero-trust systems become less reliable when they cannot distinguish an employee's actions from those taken by an agent using the same identity.

The $100 million figure covers multiple rounds. [SecurityWeek reported](https://www.securityweek.com/neo-emerges-from-stealth-with-100m-to-control-and-secure-enterprise-ai-software/amp/) that Neo received the capital through seed and Series A financings. Five days before the launch, [CTech reported](https://www.calcalistech.com/ctechnews/article/s1ukhnnvzg) that Andreessen Horowitz and Merlin Ventures had led a $25 million seed in 2025 and that Neo had recently closed more than $50 million in additional financing. Neo's launch announcement did not itemize the full $100 million by round or disclose a valuation.

### A SentinelOne operator returns to endpoint security

Warner built SentinelOne's go-to-market organization and served as president and chief operating officer as the endpoint-security vendor went public in 2021. Before SentinelOne, he held senior roles at Cylance, McAfee and Forcepoint.

Salem spent more than a decade in SentinelOne's detection engineering and threat research organization. He helped build profiles of threat actors that informed the vendor's security platform. Shirazi previously co-founded digital customer-experience company EasySend and led vulnerability research within the Israeli military's Unit 8200.

That background explains Neo's starting point. The founders are treating AI agents as endpoint software that can reason, call tools and move data while inheriting the credentials of the person they serve. An approved browser, coding tool or SaaS application can acquire autonomous functions without going through the same review process used for a new piece of enterprise software.

Warner framed the problem in operational terms: security teams need to know which software is running, what it can do, whether its configuration is safe and what should happen when it acts. Neo is selling a control plane intended to answer those questions before an agent reaches sensitive data or systems.

### Neo wants to sit between agents and their permissions

Neo says its platform inventories AI agents, AI-enabled applications, browser extensions, plugins, Model Context Protocol servers and conventional applications that have added agentic functions. A continuously updated knowledge base called Neoverse adds information about each component's capabilities, permissions, configuration and risk.

The platform also attributes actions to the human, agent, application or identity responsible. Security teams can create policies governing tool calls, API access, data movement and agent workflows for specific users or groups. Neo claims it can enforce those policies directly, including blocking risky actions or models and redirecting prompts that fall outside approved boundaries.

This positioning puts Neo across several existing security budgets. Endpoint vendors monitor activity on devices, identity vendors govern permissions, and data-security products watch sensitive information. Neo's pitch is that agent behavior cuts through all three categories because the software may be authorized, the identity may be valid and the individual action may still violate policy.

The product claims remain largely untested in public. Neo has not named customers, disclosed revenue or published performance data showing how accurately it separates human activity from agent activity. The launch materials also leave open how Neo integrates with existing endpoint and identity products, and whether enterprises will buy a separate control layer or demand these features from incumbent vendors.

Neo says it will use the financing to expand engineering and go-to-market hiring. [Israeli publication Mako reported](https://www.mako.co.il/nexter-news/Article-00ad40269fe7f91026.htm) that Neo has about 50 employees, including 40 in Israel. The large early capital base gives Warner, Salem and Shirazi room to build distribution before the endpoint, identity and data-security companies they once competed alongside fold similar agent controls into their own platforms.
