Agentic AI Workflows: MCP Integration in the SIOTH AI Engine

Department

Development department — PFE 2026–2027

Context

SIOTH® is an Industrial IT-OT-IoT platform whose AI Engine orchestrates data flows through rules and workflows driven by events, logic and scheduling. Designing these workflows requires expert knowledge.

Project

Integrate the Model Context Protocol (MCP) into the AI Engine. The MCP server will generate and manage rules and workflows in natural language, while the MCP client will add AI-agent and tool nodes to workflows.

MCP integration scope

• Server for creating, editing and explaining rules and workflows.
• Server for running, monitoring and debugging workflow executions.
• Client nodes for AI agents and LLMs inside workflows.
• Client nodes for calling external tools and enterprise systems.
• Guardrails including validation, permissions and audit trail.

Main topics

• Study MCP, AI agents and the AI Engine architecture.
• Develop the AI Engine MCP server.
• Add MCP client nodes to the workflow designer.
• Implement guardrails and human approval.
• Validate the solution on industrial use cases.

Deliverables

• MCP server and client-nodes source code.
• Sample AI-driven industrial workflows.
• Documentation including report, design report and test reports.

Expected results

Rules and workflows built from plain language, intelligent agent-driven industrial workflows, and controlled and traceable AI actions.

Keywords

MCP, AI Agents, LLM, Rule Engine, Workflow Orchestration, IIoT, .NET/C#, SIOTH.