Bringing Governance to the Age of AI Agents: Launching 'AI MONBAN' on June 9 to Visualize, Control, and Audit 'Invisible AI Risks' for Enterprises

Cyber Security Cloud, a global security manufacturer, announces the launch of 'AI MONBAN' on June 9, 2026. This AI governance platform, independently developed by its group company, DataSign, enables the visualization, control, and auditing of AI usage within enterprises.

Expanding Invisible Threats in the Age of AI and the Call for Enterprise Governance

In recent years, the business use of generative AI and AI agents has rapidly expanded, with advancements in external service integration using the Model Context Protocol (MCP). However, the use of 'Shadow AI'—AI tools not under enterprise control—is also increasing. Approximately 15% of AI users are utilizing unmanaged AI services, and one in four are inputting sensitive information into AI tools.

Guidelines issued by Japan's Ministry of Economy, Trade and Industry (METI) and the Ministry of Internal Affairs and Communications (MIC) emphasize the importance of establishing systems to appropriately monitor and control AI usage. Building a system that can explain 'who, which AI, what data, and how it was used' is now a critical management issue for companies using AI.

What is 'AI MONBAN'?

It is a governance platform that visualizes, controls, and audits all data flows related to AI usage within a company. Functioning as a 'gateway' placed between internal systems and various AI models, it allows for the centralized monitoring and control of 'who sent what to which AI, and when,' without changing existing business systems.

Main Features

- **Masking:** Automatically detects and masks confidential or personal information input into AI in real-time, preventing unintended information leaks. - **MCP Gateway:** Acts as a standard gateway compliant with the Model Context Protocol (MCP), securely connecting internal core systems and external services to AI models, enabling centralized management and flexible access control. - **AI Access Logs & Reports:** Automatically records AI usage, including data access logs via MCP and chat masking logs, to facilitate auditing and regulatory reporting without additional operational load. - **Multi-LLM Switching:** Allows users to switch between major AI models such as OpenAI, Anthropic, Google, and Azure OpenAI using only API keys, avoiding vendor lock-in and allowing selection of models optimized for specific tasks.

Future Outlook

Going forward, the company plans to support integration with existing SaaS, internal databases, and data warehouses that are not natively MCP-compatible, evolving into an integrated platform where AI agents can securely access necessary information under appropriate permissions.