How can I capture and forward user input values in an n8n AI chatbot workflow?

Below is an explanation with some guidance on how to extract values (like a budget) from the user's replies and pass them on to later nodes in your workflow.

---

Understanding the Scenario

In your workflow you have an AI Agent interacting with the user. The user's messages (for example, a reply with their budget) will be provided as part of the JSON data—in many cases available in a key such as "chatInput". Once you have that data, you can extract the specific pieces you need.

---

How to Extract the User’s Value

There are a couple of common methods:

1. Using a Set Node to Isolate the Field

   • Use an Edit Fields (Set) node to rename or copy the field containing the user's reply.
   • For example, if the incoming data has the field chatInput you could create a new field called budgetField or similar to use in later nodes.

2. Using a Code Node to Parse and Extract Information

   If you need to parse a text string (e.g., "My budget is $500 for this project."), you might write custom JavaScript in a Code node to extract the numeric value.

   For example:

   JSCopied
   // Get the user input from the incoming JSON
   const chatInput = $json["chatInput"];
   
   // Use a regular expression to capture a number following the $ sign
   const budgetMatch = chatInput.match(/\$(\d+)/);
   const budget = budgetMatch ? budgetMatch[1] : null;
   
   // Return the budget for the next nodes
   return [{ json: { budget } }];

   • This snippet extracts the budget value and makes it available in a new field named budget.
   • Ensure that the output is in the correct n8n data structure expected by subsequent nodes.

---

Best Practices for Your Workflow

• Reference Incoming Data Explicitly:
  Make sure to use dot notation (or set expressions) to reference the exact fields from the incoming data. For instance, if the AI Agent outputs its message in chatInput, then clearly reference that field.

• Clear Node Connections:
  Connect the output of your AI Agent (or your trigger node if you’re using Telegram) to the node responsible for extraction. This ensures that the data (like the user's budget) passes through correctly.

• Testing:
  You can use a Test Workflow mode in n8n to simulate the message and verify that the extraction works as expected before linking the process further.

---

Summary

You’re not facing an error message here but rather looking for a method to capture and pass on dynamic data (like a budget) from user replies. By adding a Set or Code node after your AI Agent (or Telegram trigger), you can process the incoming JSON—use expressions or regular expressions to extract the specific data you want—and then pass that data to subsequent nodes for further processing.

This approach is standard in n8n workflows that need to work with dynamic user input. The key is ensuring your nodes reference the correct field names and that the output matches the n8n data structure expected downstream.

Below is one common approach for achieving this in n8n:

1. Have the AI Agent return structured data:
   When building your prompt for the AI Agent (or any AI coding/query prompt), ask for a structured response. For example, rather than expecting a plain text reply that mentions "budget," instruct the assistant to output something like:

   PLAINTEXTCopied
   Please reply with a JSON object containing the following keys:
   {
     "budget": "<your budget value>",
     "otherInfo": "<your other response details>"
   }

   By doing this, you create a predictable structure that you can reliably reference in later nodes.

2. Extract data with a Code or Set node:
   Once your AI Agent node returns the result, you can easily extract values using dot notation. For example, with a Code node, you could do:

   JSCopied
   const items = $input.all();
   // Assuming the AI Agent returned a JSON with a key 'budget'
   const newItems = items.map(item => {
     const budget = item.json.budget;
     // You can now pass the "budget" to other nodes
     return {
       json: {
         budget,
         // include any other fields you wish to pass along
       },
     };
   });
   return newItems;

3. Pass the extracted values to subsequent nodes:
   After extracting the value(s), any following nodes can reference these properties (e.g. by using an expression like {{$json["budget"]}}) to route, transform, or use them within your workflow.

By ensuring the AI Agent produces predictable, structured output and then extracting that using Code or Set nodes, you can pass values (such as the budget) along your workflow for further processing.

If you need more detailed examples on referencing node data, see the section on referencing incoming data explicitly in the documentation.