Skip to content

Data Extraction Activity

The Data Extraction activity is a specialized component within Cognitive Services designed to retrieve structured information from a chatbot interface. By leveraging predefined Natural Language Understanding (NLU) projects, it parses input text to identify specific intents and extract the associated data for use in the automation workflow.

Field Description Requirement
Nlu project name Specifies the chatbot's project name as defined in the RPA-Admin panel. Required
Text The input string or variable containing the message from which data will be extracted. Required
Intent The specific chatbot intent name associated with the selected NLU Project. Optional
Response variable name The name of the variable where the extracted data results will be stored. Required

Action Types & Examples

Intent-Based Extraction

Extracting specific entities (like dates, names, or amounts) from a user message based on a recognized intent.

  • Format: Text: "I want to book a flight to London on June 12th."
  • Example Result: The Response variable name stores a JSON object containing identified entities like destination: London and date: 2026-06-12.

Implementation Examples

Field Setup

  • Nlu project name: CustomerServiceBot
  • Text: Global.LastUserMessage
  • Intent: BookFlight
  • Response variable name: extractionResult

Execution Parameters

  • Process: The bot sends the text to the NLU engine under the CustomerServiceBot profile and maps the recognized attributes to extractionResult.

Technical Notes

Ensure that the NLU Project is correctly published and accessible within RPA-Admin before execution. If the Intent field is left blank, the activity will attempt to identify the primary intent automatically from the provided Text.