Advanced customization of data extraction process
- Updated: 2024/12/03
Advanced customization of data extraction process
Advanced customization of extraction process enables you to modify the default extraction process when the process needs to be customized to meet your use case or requirements. By tailoring the extraction process to your specific document types and extraction requirements, you can improve the data accuracy and efficiency.
The following are some of the examples where you might need to modify the extraction process:
- Your use case requires enhancing the quality of images used for data extraction before processing the documents.
- Your use case requires modifying the extracted data before processing the data further or sending the data to system of records or user review.
When you create a learning instance, the system creates a default workflow that includes the following components in the Control Room to process documents: Automation Co-Pilot process, extraction bot, download bot, and an Automation Co-Pilot form. You can configure the form, table field options and the components of your learning instance to tune the learning instance workflow to meet your business needs.
You can do advanced customization of the data extraction process using the following process:
- Document pre-processing: Integrate pre-processing in the document processing workflow to enhance image quality, classify pages and documents before the documents are processed for data extraction. See Using pre-processing in data extraction workflow and Using classification in data extraction workflow.
- Data extraction rules: Add or modify the data validation rules, define regular expressions (regex) or patterns for data elements, and adjust the confidence levels for fields to improve data extraction accuracy. See Validation rules in Document Automation.
- Modify the default Automation Co-Pilot process: Modify this process to include custom rules or logic, integrate with enterprise systems, add custom approval workflow. See Modify default process.
- Custom parsers: Integrate your pre-trained, domain-specific models in the document processing workflow for faster deployments. See Integrate third-party parser in learning instance.