Reco has launched 'Reco AI Agent Security,' a new capability targeting the security blind spots created by AI agent sprawl in enterprise SaaS environments. Unlike traditional SSPM tools that track OAuth connections and static permissions, Reco's approach uses behavioral analysis — monitoring API call patterns, service account activity, and workflow signatures from platforms like n8n, Make, and Zapier — to detect autonomous agent behavior across multiple systems. The tool covers native agents like Microsoft Copilot and Salesforce Agentforce, and addresses real-world risks such as shadow automation with excessive permissions, misconfigured agents, and credential exposure. One cited example involved an agent exfiltrating customer data to a personal Airtable account for 8 months undetected. The capability is available immediately as part of Reco's existing SaaS security platform.
