Lumia AI Usage Control is a network-based AI security and governance platform that provides organizations visibility, control, and policy enforcement over employee and agent interactions with AI-powered applications. It operates by integrating with existing corporate network infrastructure or as a standalone proxy, using Deep Packet Inspection (DPI) and a proprietary Protocol Analysis Engine to analyze browser, native-app, and OS-level traffic for AI interactions across thousands of AI services and agents. The platform extracts user prompts, AI responses, agent actions, and metadata (user identity, privacy stance, retention policy, compliance requirements). It applies LLM and SLM-based topic analysis — defined by the organization in natural language — to classify sensitive content across text, file, voice, and visual modalities without relying on regex or manual labeling. Detected interactions are correlated with business context to enable policy enforcement actions including blocking, redacting, or logging. The platform covers standalone AI services, embedded AI capabilities within business applications, and autonomous AI agents. It is positioned for enterprise use cases where organizations need to govern what data employees and AI agents share with external AI tools, addressing data leakage, compliance, and acceptable-use policy enforcement in the context of AI adoption.