BaseMachina Opens Pre-Registration for “BaseMachina AI SecureOps” to Add Governance to AI Agent Operations on Production Data

BaseMachina Inc. (Chuo-ku, Tokyo; CEO: Seiji Takahashi), which has provided business infrastructure such as internal admin panels for web companies as well as systems needed for product operations and internal workflows, announced that it has opened pre-registration for “BaseMachina AI SecureOps.” The service is designed to add governance mechanisms to AI agent operations on production data, including permission management, approval workflows, audit logs, and notifications. BaseMachina AI SecureOps is a service for adding governance to workflows in which AI agents access production databases, data warehouses, internal APIs, and core business systems. By connecting to existing databases or APIs, the service plans to automatically generate remote MCP servers that AI agents can use. For workflows that do not require an admin screen, users will be able to execute business operations through MCP from AI tools they already use, such as Claude. For workflows that require human review or operation through a screen, BaseMachina also plans to enable coding agents such as Claude Code and Codex to generate admin UI based on CLI commands and configuration files. The company aims to let organizations operate AI-facing tools and human-facing admin screens under the same definitions and the same governance model, rather than managing them separately. Planned key features include automatic generation of remote MCP servers that turn SQL queries and API calls to production databases, DWHs, and internal APIs into business tools usable by AI agents; CLI and agent skills that support AI-driven configuration generation, allowing coding agents to generate action definitions and configuration files based on data sources and business requirements and reflect them in the service; GitHub-based management and review of generated configurations, including history tracking, diff review, approvals, and rollback; tool-level permission management that controls executable scopes by operation type such as read or write, as well as by role or agent; approval workflows that insert human approval for risky operations such as refunds, status changes, and updates to important data; audit logs that record who executed which tool, when, and through which AI agent for audits and incident investigations; notifications for approval requests and execution results via Slack, email, Webhook, and other channels; and automatic admin UI generation from the same configuration files when human review or operation screens are needed. BaseMachina said it has historically provided screens and workflows needed for internal operations and product management as an admin panel development platform. The company noted that customers valued not only the ability to quickly build admin screens, but also governance features required for safe operations, such as approval workflows, business-specific permission management, audit logs, and notifications. The company believes AI agents will increasingly become actors that execute business operations and access production product data and internal systems. As a result, it argues that expanding AI usage is not enough; companies must also create conditions in which AI can perform business operations safely. With BaseMachina AI SecureOps, the company plans to apply the governance capabilities it has developed through its admin panel platform to AI-agent data access and data updates. Through this pre-registration, BaseMachina will gradually provide information on beta and general availability to companies that want to safely connect AI agents to production databases, DWHs, administrator APIs, core systems, and similar environments. The company says it will continue developing AI SecureOps as a product that supports data governance in the AI era, aiming not only to bring AI into business operations but to build the foundation for AI to perform those operations safely.