Every enterprise rushing to deploy AI agents is also, whether they realize it or not, punching new holes in their security perimeter. Reco, an AI-native security platform built to monitor and protect those agents alongside SaaS applications, just raised $50 million in a Series B round to make sure someone is watching. The funding, Globes reported, was led by Advent International, with participation from existing investors Insight Partners and Boldstart Ventures.
The raise puts Reco in a crowded but rapidly expanding category — one where the threat surface is growing faster than defenders can map it. As more organizations hand autonomous agents access to email, code repositories, CRM platforms, and internal databases, the question of what those agents are actually doing, and whether they can be hijacked, poisoned, or abused, has become urgent. It is a problem closely related to the agent control failures that have already surfaced at some of the industry’s largest AI labs.
What Reco Actually Builds
Reco’s platform focuses on identity and access visibility across SaaS environments, tracking who and what has access to which systems, and flagging anomalies that suggest compromise or misuse. With AI agents entering those same environments, the company has extended its monitoring capabilities to cover non-human identities — bots, pipelines, and autonomous software that can be granted sweeping permissions across an organization’s cloud stack without the usual human oversight checks.
The company’s technology maps out the relationships between users, data, and applications in real time, building a behavioral baseline that lets it detect when something is operating outside expected parameters. That matters especially in agentic workflows, where a compromised or misbehaving agent might exfiltrate data, escalate its own privileges, or pivot laterally across connected services before a human analyst notices anything unusual. According to Globes, Reco’s platform is already deployed across hundreds of enterprise customers, giving the company a meaningful data advantage as it trains its detection models.
Why $50 Million, Why Now
The timing of this round is no accident. Enterprise adoption of AI agents has jumped sharply in 2025 as major SaaS vendors, from Salesforce to Microsoft, have embedded agentic capabilities directly into their core products. That integration is convenient for productivity and catastrophic for traditional perimeter-based security models, which were never designed to handle software that can initiate actions, authenticate itself, and chain together complex multi-step tasks without a human in the loop.
Reco is not alone in chasing this market. Competitors including Valence Security and Astrix Security are also building identity and integration security tools aimed at SaaS and AI environments. But Reco’s Advent-led raise, at a moment when enterprise security budgets are under scrutiny, signals genuine investor conviction that the AI agent security layer is a durable, mission-critical product category rather than a passing feature wedge. The round is comparable in scale to what peers in adjacent AI security markets have recently commanded — similar to the $50 million Unframe secured for governed AI integration infrastructure earlier this year.
With the fresh capital, Reco says it plans to accelerate product development and expand its go-to-market team, particularly in North America, where enterprise demand for AI governance and security tooling is concentrated. If the agent deployment wave continues at its current pace, the window to become the default security layer for that infrastructure is narrow — and Reco is moving to close it.
