Extraction Activity

The Extraction activity is part of the Document Understanding suite, designed to pull structured data from various document types. It utilizes specialized models trained within the RPA-Admin interface to recognize and export specific fields from digital or scanned images.

Field	Description	Requirement
Document type	Specifies the project or document category trained in the RPA-Admin Document Understanding section.	Required
Image file name	The full file path of the document image (JPG, PNG). PDF documents are not supported.	Required
Document language	The language of the text within the document (e.g., "tur", "eng").	Optional
Response variable	The variable name where the extracted structured data will be stored as a result.	Required

Action Types & Examples

Structured Data Extraction

Retrieving key-value pairs (e.g., Invoice Number, Date, Total) from a specific document class.

Format: Document type: "Invoice_Model_V2"
Example Result: {"InvoiceNo": "INV-2026-001", "Date": "17.03.2026", "Total": "1500.00"}

Implementation Examples

Field Setup

Document type: TR_ID_Card
Image file name: C:\Scans\identity_front.jpg
Document language: tur
Response variable: extractedIDData

Execution Parameters

Process: The system sends the image to the Document Understanding engine. Using the TR_ID_Card model, it identifies the relevant fields and maps them to the extractedIDData variable.

Technical Notes

The available options for the Document type field are directly synchronized with the projects created and trained in the RPA-Admin Document Understanding module. Ensure that the model is published and that the Document language matches the source text to optimize extraction accuracy.