| Term | Description |
|---|---|
| Model Context Protocol (MCP) | A standardized protocol that enables AI agents to securely interact with external data sources and services. The protocol provides a framework for tool discovery and execution through JSON-RPC 2.0 communication. |
| Agent Tools (MCP) Asset | A Celonis asset type that encapsulates a collection of ready-to-use tools that AI agents can use to search for relevant information, get real-time process context, trigger actions in external systems, and write back decisions to Celonis. It provides access to Celonis data and insights through standardized JSON-RPC 2.0 protocol endpoints and OpenAPI tool-calling endpoints. |
| MCP Client | An application or service that connects to an MCP Server to discover and execute tools. Examples include Claude Desktop, Postman, and other MCP-compatible clients. |
| Tool | A function that AI agents can call through the MCP Server to perform specific tasks such as searching data, retrieving process insights, or triggering actions. Each tool has a unique name, description, and input/output schema. |
Last updated