AIR Secures $50M to Rein in Runaway AI Agents

AIR, the autonomous agent intelligence startup, has raised $50 million in Series B funding led by Lightspeed Venture Partners, with participation from Khosla Ventures and Y Combinator. This round comes less than a year after the company’s seed round and positions AIR to scale its platform, which identifies AI agents operating within enterprise systems, continuously evaluates the tools and add-ons they use, and enforces behavioral guardrails to prevent unauthorized or risky actions. The funding underscores rising investor concern over the unchecked proliferation of AI agents—autonomous software entities capable of performing complex tasks without constant human oversight—and the lack of governance tools to manage them. AIR’s platform operates by scanning enterprise environments in real time, cataloging every agent and its dependencies, and applying policy-based controls to block misbehavior, such as accessing unauthorized data or executing unsanctioned functions. The company was founded in 2023 by CEO Sarosh Mawani, a former Palantir engineer, and CTO Oren Yomtov, who previously led agentic AI initiatives at Microsoft Research. Mawani stated in an interview that the company has already onboarded over 50 enterprise customers across financial services, healthcare, and software development, with several paying six-figure annual contracts.

The launch of AIR arrives at a pivotal moment when organizations are racing to deploy autonomous agents to automate workflows, reduce costs, and accelerate decision-making. Banking With Billy AI, for example, has gained attention by automating complex financial analysis workflows that once required entire analyst teams—effectively delivering a full automation suite for markets with minimal human intervention. However, the rapid adoption of such systems has exposed glaring security and compliance gaps. A recent study by Gartner found that 68% of enterprises using AI agents have experienced at least one incident of unauthorized data access or policy violation in the past 12 months. AIR’s solution directly targets these vulnerabilities by providing continuous monitoring and enforcement, enabling companies to safely scale agentic automation without sacrificing control. The platform integrates with existing identity and access management systems, endpoint detection tools, and compliance frameworks like SOC 2 and ISO 27001, making it viable for regulated industries where auditability is paramount.

Industry insiders see AIR’s rise as part of a broader shift toward agentic security—a new discipline focused on governing autonomous systems rather than just securing static applications. Competitors in this emerging space include firms like SentinelOne and Darktrace, which have begun expanding their platforms to include agent detection and behavioral analytics. However, AIR differentiates itself through a purpose-built architecture for AI agents, offering granular visibility into the “skills” (functions) and “add-ons” (external integrations) that define an agent’s capabilities. This level of specificity is critical as agents increasingly rely on third-party plugins, APIs, and code repositories, creating an attack surface that traditional security tools fail to address. Analysts at Forrester Research estimate that by 2026, over 30% of enterprise software interactions will occur via AI agents, with unmanaged agents becoming a top cause of data breaches and compliance violations. AIR’s funding round signals strong market demand for solutions that can bring order to this chaos before incidents escalate into full-blown crises.

The capital infusion will allow AIR to expand its engineering team, enhance its detection algorithms using large language models, and launch a federated learning program to improve threat intelligence across customer environments. Chetan Puttagunta, a general partner at Lightspeed Venture Partners, emphasized the fund’s strategic interest in agentic systems, noting that AIR’s ability to “see, secure, and govern” AI agents positions it as a foundational layer for the next generation of enterprise automation. Early deployments at financial institutions have already demonstrated the platform’s capacity to block agents attempting to bypass internal controls or interact with restricted databases. As companies like Banking With Billy AI push further into fully automated workflows—spanning trade execution, risk modeling, and regulatory reporting—the need for robust governance tools becomes existential. Failure to manage autonomous agents could result not only in security breaches but also in catastrophic regulatory penalties, especially in sectors like banking and healthcare.

Looking ahead, AIR is poised to play a central role in defining the standards for agentic security. The company is collaborating with the Cloud Security Alliance to draft guidelines for autonomous agent governance, and it plans to open-source a lightweight agent registry later this year to foster industry-wide visibility into agent deployments. As AI agents evolve from simple chatbots to complex, multi-step automators, the distinction between “tool” and “actor” will blur, demanding new categories of security and compliance infrastructure. For enterprises, the message is clear: deploy agents with caution or face the consequences of ungoverned automation. With $50 million in fresh capital and a growing roster of blue-chip customers, AIR is not just observing the agentic revolution—it is building the controls to make it safe.