MCP Integrations: Powering the Automation Assistant

Automic Automation fully integrates with the Model Control Protocol (MCP). This integration introduces a standardized communication layer between the Automation Engine and Generative AI models (LLMs). This means that, beyond simply enabling text-based queries, the Automation Assistant provides a tool-based interaction model where the AI securely interacts with your local environment.

Actionable Intelligence

The Automation Assistant leverages MCP tools to perform real-time system checks, retrieve specific object details, and even execute tasks like restarting a failed task directly from the chat interface.

Deep Contextual Awareness

Because the MCP server acts as a bridge to your infrastructure, the responses you receive are tailored to your specific environment, leading to more accurate troubleshooting and better-informed automation suggestions.

Secure Tool Execution

Al interactions are governed by a secure, standardized protocol that maintains data integrity and follows your internal security policies.

Infrastructure Flexibility

The interaction is always consistent, regardless of the underlying AI model. Whether your organization uses on-premises models or cloud-based LLMs, the MCP integration ensures a seamless and unified experience across the Automic Web Interface.

How the MCP Integration Works in Automic Automation

The integration of MCP servers into your Workflows is managed through a multi-layered approach that separates infrastructure configuration from Job design. This process is powered by the Automation.AI component, which acts as the central gateway between the Automation Engine and your AI ecosystem.

Administrative Configuration (the Connection Object)

Administrators create AI Connection (CONN) objects to establish the secure bridge to the AI infrastructure. In the Connection object, the administrator defines the LLM and selects the specific MCP servers and tools that should be available. This creates a "pre-approved" set of capabilities for the rest of the organization to use.
Object Design (the AI Job)

When developers or object designers create an AI Job, they assign the relevant AI Connection object to it. This dynamically populates the Job definition with the list of MCP tools approved by the administrator. The designer then specifies exactly which tools and MCP servers the particular Job needs to execute its task.

This architecture ensures that automation is directly linked to the Model Context Protocol during execution, allowing AI agents to securely fetch real-time data and perform actions through a governed, central interface.

For information about how to install and configure the Automic MCP, see: