AI Agent Workflow: Cold Outreach Subworkflow Not Receiving Data

Below is some guidance on both troubleshooting the data not arriving at the cold outreach subworkflow and on how to include code using JSON markdown formatting.

---

Troubleshooting the Data Transfer Issue

1. Verify Input Mapping in the Subworkflow
   Check that the subworkflow’s parameters (e.g. in the “Edit Fields” or similar nodes) are correctly set up to capture the incoming data. In many cases the data isn’t “lost” but isn’t mapped to the expected variable. Make sure that the field name you’re expecting (for example, the Sheet_ID or other values) matches the key in the request sent from the AI agent.

2. Check the Custom n8n Workflow Tool Configuration
   When using the Custom n8n Workflow Tool (used by the agent to call a subworkflow), verify that the “fields” section is correctly defined so that it passes the data you expect. Ensure that any expressions (using {{ $json.someField }} or similar) are returning the correct values before the request is sent.

3. Review the Data Route
   Since you mention that the execution log shows data in the request, compare the sent payload with the subworkflow’s expected input. You might want to add a debug node (for example, a “Set” or “NoOp” node that simply logs the incoming data) at the very beginning of the cold_outreach_message tool subworkflow. This helps in checking if there’s any transformation or naming discrepancy.

4. Execution Order and Wait-Conditions
   Confirm that you do not have any race conditions or unexpected execution orders. Since the AI agent is designed to run tools sequentially (waiting for a “done” response), ensure that the subworkflow is receiving the request at the correct time and that no step early in the subworkflow is dropping or overriding the incoming data.

5. n8n Version and Settings
   You’re running on version 1.78.0 with SQLite and the EXECUTIONS_PROCESS setting set to "main." While these should work, verify in your logs whether any warnings or errors are (even intermittently) being reported regarding data structure or connections between workflows.

---

Including Code in JSON Markdown Format

Your use of triple backticks for enclosing JSON code is correct. For example, using:

JSONCopied
{
  "nodes": [
    {
      "parameters": {
        "promptType": "define",
        "text": "=You are a helpful assistant for Lead Generation ...",
        "options": {}
      },
      "type": "@n8n/n8n-nodes-langchain.agent",
      "typeVersion": 1.7,
      "position": [440, 0],
      "id": "dd2311bb-4cd4-4851-8772-b44b22f3523b",
      "name": "AI Agent"
    },
    { ... }
  ],
  "connections": { ... }
}

is perfectly acceptable. Just ensure that the JSON is valid (for example, using straight quotes rather than “curly” ones, as some editors may auto-format them) so that others can easily copy and paste the configuration if needed.

---

Summary

• Data not reaching the subworkflow? Verify that the incoming fields are correctly mapped in the subworkflow. Add a debug/set node near the start of the subworkflow to inspect the received payload.
• Custom Workflow Tool: Double-check the configuration especially the fields section where you reference earlier data (like Sheet_ID) and ensure the expressions return the intended value.
• JSON Code Blocks: Your approach using triple backticks with “json” is recommended. Just make sure the JSON syntax remains valid (straight quotes, proper commas, etc.).

Try these steps and check your execution logs again to narrow down where the data isn’t matching the expected variable names. Let me know if you need more specific assistance on any of these points.

Below is some guidance that may help you troubleshoot and resolve the issue:

1. Verify How Data Is Passed to the Subworkflow

When you call a subworkflow (or tool workflow) from your AI Agent node, n8n passes the data as part of the request payload. In your case, the execution log shows that data is being sent but the receiving subworkflow (the one containing your cold_outreach_message tool) isn’t picking it up. Here are a few points to check:

• Trigger Node in the Subworkflow:
  Make sure the subworkflow’s entry point (for example, a Workflow Trigger or Webhook node) is set up to receive incoming data. The data you pass must be handled by that node. If you’re using a Workflow Trigger, check that the incoming JSON is available (via the “$json” object) to subsequent nodes.

• Field Names and Data Mapping:
  Ensure that the field names you pass from your AI Agent match what the subworkflow is expecting. For instance, if the cold_outreach_message tool is expecting a variable named Sheet_ID (or any specific key), confirm that:

  • The AI Agent sends that key.
  • The subworkflow references it correctly (for example in expressions like {{ $json.Sheet_ID }}).

• Pass Input Data Option:
  Some n8n nodes (especially custom tool workflows) require you to explicitly enable “pass input data” or similar options. Double-check that in your AI Agent’s tool call configuration (and in the subworkflow’s settings) you haven’t inadvertently disabled data passing.

2. Check Your JSON Formatting

Your example shows code wrapped in triple-backtick code blocks marked as “json.” This is generally acceptable for documentation and debugging purposes. However, note these points:

• Use Standard Double Quotes:
  Ensure your JSON uses standard double quotes ("), not typographic quotes (“ ”). Typographic quotes may cause parsing issues when n8n reads your configuration.

• No Additional Wrapping Needed:
  When importing or editing nodes in n8n, you usually paste raw JSON (using native quotes) directly into the n8n Editor or load them via the Import function. The markdown triple-backticks are only for presenting the code. In your n8n instance, ensure that the code field is exactly as expected.

3. Debugging Steps

• Check the Execution Log:
  In n8n’s execution details for the subworkflow, inspect the JSON payload received at the entry node. This will confirm whether the data is arriving as expected.
• Test with a Simple Value:
  Temporarily configure your subworkflow to simply log or pass forward the entire incoming JSON (using a “Set” node without alterations). This can help you verify that the data is reaching the subworkflow.
• Review Tool Workflow Configurations:
  Make sure that in your AI Agent’s tool call (the cold_outreach_message tool), the settings for “workflowId” and the fields mapping are correct. A common mistake is to have a mismatch between what is sent and what the subworkflow expects.

Summary

• Confirm the subworkflow’s trigger node is correctly set up to receive data.
• Check that the field names between the AI Agent and subworkflow are consistent.
• Make sure your JSON uses standard quotes (") when configuring nodes.
• Use n8n’s execution logs and “Set” node tests to verify incoming data.

By following these steps, you should be able to pinpoint whether the issue is due to misconfigured data mapping in the tool call or a formatting issue with the JSON itself.