Model Context Protocol (MCP) tools in Planning Analytics

Model Context Protocol (MCP) tools make Planning Analytics data and analytical capabilities available to agentic workflows and AI-enabled assistants.

Overview

Model Context Protocol (MCP) tools provide structured, callable capabilities that allow external agents and AI-enabled services to interact with Planning Analytics systems. These tools are designed for use in agentic workflows, such as Planning Analytics Agent and other MCP-enabled integrations.

These tools focus on providing intent-level capabilities rather than implementation-level APIs. Agents use natural-language requests or structured calls to invoke MCP tools, and these tools return results that agents use to reason about data, metadata, and analytical outcomes.

Additionally, all MCP tools are accessed through a single endpoint, /ibm-pa-tools. This unified endpoint replaces earlier endpoints that were used for specific tool groups. Tool groupings such as cube tools and analysis tools are no longer exposed as separate endpoints. You can access all tools from the unified endpoint and manage tool selection within your application.

MCP tools can also retrieve metrics, including resource usage and data statistics, for TM1 12 databases, cubes, and replicas. Metrics tools require a version of TM1® that supports the Metrics API (version 12.5 or later).

Scope and purpose

MCP tools are intended to support exploration, analysis, and interpretation of Planning Analytics models. These tools support agentic workflows by enabling agents to perform the following functions:

  • Discover available Planning Analytics resources.
  • Inspect cube structures and metadata.
  • Retrieve data and analytical results.
  • Generate context-aware insights and summaries.

These tools are designed for technical audiences and integrations that require programmatic or agent-driven access to Planning Analytics capabilities.

Relationship to agentic workflows

MCP tools do not operate independently. Agents invoke these tools to interpret user intent, identify relevant tools, and assemble responses based on tool outputs.

The agent or consuming service selects and orchestrates the tools, while Planning Analytics provides the tools themselves. Agents determine how and when to use them.

Tool families

For clarity, Planning Analytics documentation groups MCP tools by function based on the capabilities that they provide.