Neo emerged from stealth on Monday with $100 million in funding led by Andreessen Horowitz and Bessemer Venture Partners, with participation from Craft Ventures and Merlin Ventures. The company is building what it calls a real-time control layer for agentic software in the enterprise, addressing a critical and fast-emerging security gap as artificial intelligence agents become embedded into browsers, developer tools, SaaS platforms, and legacy applications at a pace that outstrips traditional security monitoring.
The funding round, unusually large for a seed-stage company, signals that investors see a massive market opportunity in managing the security implications of autonomous software. The problem Neo solves stems from a fundamental shift in enterprise software architecture: applications that once operated deterministically—following predefined rules and responding only to explicit user commands—are now gaining the ability to reason, act, invoke external tools, chain complex workflows, and inherit user permissions. This transformation, driven by advances in large language models and agentic frameworks, is happening inside environments built for predictable, human-directed applications.
The Scale of the Shift
According to Gartner, only 5% of enterprise applications had agentic capabilities in 2025. By the end of 2026, that figure is projected to jump to 40%. This means that within the next 18 months, nearly half of all software used by large organizations could be capable of autonomous decision-making and action-taking. The security implications are profound: each agentic application represents a potential vector for unintended data exposure, unauthorized access, or cascading failures if the agent's actions are not properly constrained.
Traditional security tools were designed to monitor human users and static applications. They can log what a user clicks or what data an application accesses, but they are largely blind to the internal reasoning and autonomous actions of AI agents. An agent that is granted permission to read a database may decide to write new entries, invoke a third-party API, or share data with another agent—all without explicit human approval or oversight. Neo aims to fill this visibility and control gap.
Neo’s Platform: Inventory, Scoring, and Enforcement
Neo’s platform gives security operations teams an inventory of every AI agent and agentic-enabled application running across the organization. It scores each agent’s capabilities and associated risks, providing a clear picture of what each piece of autonomous software can do. The platform also maps actions back to the responsible user or agent, creating an auditable trail of decisions and behaviors. Most importantly, Neo enforces policy before risky activity occurs, blocking dangerous tool calls, API access, or data movement at the point of action.
The enforcement is native, meaning Neo intercepts and controls agent behavior without routing traffic through another security product. This architectural choice reduces latency and complexity, allowing security teams to apply granular policies directly to agentic workflows. For example, an agent attempting to access a restricted database or invoke a sensitive API can be stopped in real time, with a clear explanation of why the action was blocked. This proactive approach contrasts with reactive monitoring, which can only report incidents after they have occurred.
Founders with Deep Security DNA
Neo was founded by executives with experience at some of the most prominent companies in cybersecurity. CEO Nick Warner previously designed SentinelOne’s go-to-market operations and played a key role in taking the company public in 2021. SentinelOne, a leader in endpoint detection and response, built its reputation on autonomous AI-based threat hunting—giving Warner firsthand insight into how AI can both defend and challenge enterprise security.
The founding team also includes alumni from Wiz, a cloud security platform that grew rapidly after its founding and was recently acquired by Google for $32 billion, and Palo Alto Networks, the long-established network security giant. This mix of backgrounds—endpoint security, cloud security, and network security—positions Neo to address agentic AI risks across multiple layers of the enterprise stack. The team’s collective experience in building scalable security products that operate in real time was a key factor in attracting top-tier venture capital.
Competitive Landscape
Neo enters a quickly expanding category of startups focused on agentic security. Earlier this year, Straiker raised $64 million for a similar security platform, and NewCore raised $66 million to give AI agents corporate identities, focusing on the authentication side of the problem. While Straiker and NewCore address complementary aspects, Neo is betting that the bigger market opportunity lies in the control layer: not just who the agent is (authentication) but what it is allowed to do (authorization and enforcement).
This distinction is crucial because agentic AI operates differently from traditional software. An agent may be authenticated as a legitimate entity but still engage in activities that violate security policies through chains of reasoning or tool usage that were not anticipated by its developers. Neo’s approach focuses on behavioral control, ensuring that even authenticated agents can only perform actions that have been explicitly approved.
Background and Market Context
The rise of agentic AI is not limited to a single software category. Major enterprise vendors, including Salesforce, Microsoft, and ServiceNow, are embedding autonomous capabilities into their platforms. These features are often added to products that are already approved for internal use, meaning security teams may not even be aware that a trusted application has suddenly gained the ability to act on its own. This is the core challenge Neo addresses: the risk is not from rogue AI tools that employees download without permission, but from approved software that quietly becomes autonomous.
Agentic AI applications include browser extensions that can fill forms and navigate websites, developer tools that can write and deploy code, SaaS platforms that can automate workflows across multiple services, and legacy applications that gain agentic features through updates. Each of these creates new attack surfaces. For instance, a browser extension with agentic capabilities could read all emails and calendar events, then use that information to take actions on the user’s behalf—potentially exposing sensitive data or making unintended changes.
Technical Depth: How Enforcent Works
Neo’s platform operates by integrating with the agent’s runtime environment, whether that is a browser, a developer environment, or a cloud-based API. It monitors every tool call, every data access request, and every workflow chaining operation. The control layer uses a policy engine that administrators can configure with rules based on agent identity, capabilities, context, and risk scores. Enforcement decisions are made in milliseconds, ensuring that agent productivity is not significantly impacted while security is maintained.
The platform also provides a dashboard that aggregates data from all monitored agents, giving SecOps teams a unified view of autonomous activity across the enterprise. This includes real-time alerts for policy violations, historical logs for forensic analysis, and risk trend reports that can help prioritize which agents to investigate or restrict. Neo claims that its native enforcement approach avoids the overhead of redirecting traffic through a separate proxy, which can slow down agent operations and increase infrastructure costs.
CEO Perspective
“Enterprise security was built for a world where software behaved predictably. That world is changing fast,” said Nick Warner, CEO of Neo. “We are seeing an unprecedented explosion of autonomous capabilities inside software that was already trusted and deployed. Our platform gives security teams the visibility and control they need to embrace agentic AI without compromising security.”
Warner’s comment underscores the dual challenge: enterprises want to adopt AI agents because they boost productivity and enable new workflows, but they cannot afford to let these agents operate without oversight. Neo aims to provide the missing safety layer, allowing companies to move quickly while keeping risk in check.
The Investment Signal
At $100 million in seed-stage capital, a16z and Bessemer are making a bold bet that Neo can define and dominate the agentic security category. The amount is unusual for a seed round, reflecting both the urgency of the problem and the conviction that the winner in this space will have a significant first-mover advantage. The investors are betting that whoever builds the control layer first—and earns the trust of enterprise security teams—will own the category for years to come.
The timing aligns with broader market trends. As agentic AI moves from experimental to production in sectors like finance, healthcare, and technology, the need for specialized security controls becomes acute. Regulation is also likely to play a role: frameworks such as the EU AI Act and emerging U.S. guidelines may require organizations to demonstrate that autonomous systems are subject to human oversight and policy enforcement. Neo’s platform could help meet these compliance requirements by providing an auditable control layer.
Looking ahead, Neo plans to expand its platform to cover more types of agentic software and to integrate with existing security information and event management (SIEM) systems. The company also intends to build out its partner ecosystem, working with cloud providers and enterprise software vendors to ensure broad compatibility. The $100 million will fuel aggressive hiring in engineering, sales, and customer success as Neo works to establish itself as the standard for agentic AI security.