### Answer Based on the documentation, the recommended approach is to include explicit instructions about when to invoke the MCP tool in the prompt itself. [(1) ]() The documentation states: > "It's best to include in the prompt explicitly when is the best time to invoke the mcp tool." With an example like:

When user states his name and phone number, please call tool `verify_user` function.

For Conversation Flow agents specifically, you have multiple levels where prompts can be configured: - Global Prompt (

global_prompt

) : Used in every node of the conversation flow [(2) ]() [(3) ]() - **Node-level instruction**: Each conversation node has its own instruction/prompt The sources don't explicitly specify whether the MCP invocation instruction should go in the global prompt versus the node-level prompt for Conversation Flow agents. However, the general guidance suggests placing the instruction where it's most contextually relevant to when the tool should be called. A practical approach would be: - Use the node instruction (the conversation prompt within or before the MCP node) if the MCP tool is only relevant to a specific part of the conversation - Use the global prompt if the MCP tool might be needed across multiple nodes Check out these potentially useful sources for more details: - [MCP]() - [Retell AI's Advanced Conversation Flow]() Would you like a general answer based on my knowledge about prompt placement best practices?

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

​ To learn more, see the sources I considered: [[1] MCP](https://docs.retellai.com/build/single-multi-prompt/mcp) [[2] Get Conversation Flow](https://docs.retellai.com/api-references/get-conversation-flow) [[3] Update Conversation Flow](https://docs.retellai.com/api-references/update-conversation-flow) ​