Reco Enhances AI Runtime Security with Browser-Based AI and Automated Remediation

Post Views: 10

Reco has introduced enhancements to AI Runtime, a foundational element of the Reco Platform, incorporating browser-based enforcement, real-time analysis, and automated resolution processes.

Enhancements to AI Runtime

AI Runtime is designed to manage risks in real time without requiring security teams to reroute traffic through additional gateways or proxies. Each deployed agent within an organization presents a potential attack surface, defined by the applications it can access, the permissions it holds, and the connections it initiates. Reco currently maps this attack surface.

Browser Guard and Smart Remediation

\\\”Posture defines the scope of what an agent can access, whereas runtime enables control over its current actions without necessitating traffic redirection through a gateway,\\\” explained Ofer Klein, CEO of Reco. \\\”Browser Guard secures the environment where most personal AI interactions occur.\\\”

Smart Remediation, released in July, transforms risk assessments into automated corrective actions, completing the cycle of identifying, prioritizing, and resolving threats across the Reco Platform. The system evaluates the specific requirements of each agent to avoid disrupting critical workflows when addressing risks.

Reco Browser Guard Details

Reco Browser Guard, currently in early access, provides visibility into AI usage outside approved channels, focusing on browser activity where shadow AI practices typically originate. The initial phase identifies such usage by tracking which employees utilize specific AI tools, the associated accounts, and frequency of access. A second phase will introduce browser-level blocking and policy enforcement, scheduled for early access on July 29.

Real-Time Analysis and Policy Engine

Real-time analysis of prompts, responses, and tool interactions, set for general availability in October, will extend Reco’s policy engine across all stages of the agent layer. This process examines each prompt, AI-generated response, and tool call against the Reco Graph, offering insights into the involved agent’s risk profile, connected systems—including all MCP servers it interacts with—and the user’s broader SaaS and AI ecosystem.

Conclusion

Reco’s updates aim to address evolving challenges in securing AI-driven environments while minimizing operational disruptions. The platform’s focus on real-time risk mitigation and automated responses reflects growing demands for proactive security measures in enterprise settings.