How can I force AI-Agent to call tools serially, and not in parallel?

Questions
1

I currently have an AI-Agent, that has another AI-Agent as a tool, connected to the new Data-Tables MCP. I have found a problem whereas if the Data-Tables MCP tries to access the Data-Table more than once concurrently, only the first access succeeds. So as a workaround how can I get things such that the main AI-Agent only queries the AI-Agent tool serially, waiting for a response before asking for next item?

2

To force serial execution of AI Agent tools, you can:

1. Use the **Execute Once** option in the tool node’s settings to process one item at a time ([docs.n8n.io]( Looping | n8n Docs ))

2. Alternatively, implement a callback pattern where the main workflow waits for each tool call to complete before proceeding ([community.n8n.io]( Issue with parallel pattern when using AI Agent tool call ))

For Data Tables access specifically, consider adding a small delay between operations or implementing a queue system.