Fathym has expanded Open Industrial to let any MCP-compatible AI tool propose governed workflows, bringing vendor-neutral AI interoperability and human-approved automation to industrial and regulated environments.
Fathym has announced that its Open Industrial platform now allows any MCP-compatible AI tool to propose workflows directly into governed enterprise environments, expanding open AI interoperability for industrial and regulated sectors.
The platform’s expanded MCP server supports external AI tools including Claude Code, GitHub Copilot, custom AI agents, and any frontier model accessible through the Model Context Protocol. Developers can work from their preferred coding environment, generate workflow proposals in natural language, and submit them directly into Open Industrial for approval.
The company stressed that no AI-generated workflow executes automatically. Every proposal must pass through human review, explicit approval, governance checks, attribution logging, and audit tracking before deployment.
Open Industrial also introduced an API Source node that connects REST endpoints as live data feeds without custom integration code. The feature supports polling intervals, authentication, response mapping through a visual inspector, a manual “Pull Now” validation mode before deployment, and continuous endpoint health monitoring.
According to the company, the release addresses a long-standing enterprise AI challenge: balancing developer freedom with governance and compliance requirements. Open Industrial said its bidirectional MCP architecture enables organisations to use preferred AI tools while maintaining oversight and auditability.
“Organizations no longer need to choose between the AI tool they want and the governance they require. Any AI tool that speaks MCP can now propose work into a governed environment — and nothing it proposes executes without a human in the loop. That is the architecture teams have needed,” said Matthew Smith.
