Validating documents overview
- Updated: 2024/09/17
Validating documents overview
In many business scenarios, data extraction from documents is a critical part of automation. While extraction processes are designed for high accuracy, the reality is that documents can vary significantly in structure, format, and quality.
For example, invoices might be formatted differently, documents might be scanned or handwritten, and contracts often deviate based on their origin. These variations introduce complexities that can impact the accuracy of extracted data.
The validation feature addresses this by acting as an additional safeguard, ensuring that any inconsistencies or errors caused by document variability are identified and corrected. This is especially important in industries with strict compliance requirements, where even a minor data error could have significant legal or operational consequences. By validating the extracted data, businesses can ensure its accuracy and reliability before it moves into downstream processes.
Overview
The system identifies documents that require validation through several mechanisms. By default, a document is flagged for validation based on the following criteria:
- Missing required fields
- In-built data type validations (for example, number and date)
- Fields with low confidence
- Custom validation rules defined by the user
There are two types of validation flags:
- Errors: Documents with errors must be corrected before submission.
- Warnings: Documents with warnings can be submitted without changes.
For example, a low-confidence field will trigger a warning, while a missing required field will generate an error. When creating custom validation rules, it is important to assign the appropriate flag based on the severity of the issue.
In addition to the default checks, users can define workflow rules that route documents to the validation queue as per their requirements.Validation UI
To assist with the validation process, Automation Anywhere provides a user-friendly interface through the Automation Co-Pilot feature. This UI clearly highlights errors and warnings in the document, allowing users to easily identify and address issues. After all validation errors or warnings are resolved, the document is successfully validated and can proceed for further processing. In addition to the validation UI, Automation Co-Pilot offers robust validation queue management. Users can view all validation tasks in a consolidated view, filter tasks based on criteria, search for specific tasks, and assign tasks to different users. This ensures efficient task management and enables teams to handle validation at a scale.
How validation works
You can access the Automation Co-Pilot validator in the following ways:
- Log in to the Control Room as a validator user. On the Learning Instances page, click the Validate documents link that has pending document validations. See Validate documents as Validator user.
- Log in to Automation Co-Pilot as a Automation Co-Pilot admin and click Tasks. From the Tasks page, you can assign, filter, and search validation tasks. See Assign or unassign a task and Filter and search for a task.
To understand how validation works, see Validating documents through Automation Co-Pilot validator.