AI-Powered Document Record Extraction in Oracle Cloud HCM

In the fast-paced world of HR administration, data entry is a notorious bottleneck. Between handling employee certifications, visas, identification papers, and medical records, HR professionals and managers spend hours manually transcribing data from physical documents into Oracle Cloud HCM.

Manual data entry isn’t just time-consuming—it is also prone to human error.

With the Oracle Cloud HCM 26C release, Oracle is tackling this challenge head-on. Introducing AI-Powered Document Record Extraction, a feature designed to automate manual tasks and introduce a new level of efficiency to HR document management.

Let’s dive into how this feature works, the tangible business benefits it brings, and how you can enable it for your organization.

What is AI-Powered Document Record Extraction?

When creating a new document record on the Redwood interface, users can now upload an attachment (such as a PDF or image) and let Oracle’s embedded AI do the heavy lifting. By clicking an “Extract” button, the system automatically scans the attachment and prefills standard fields.

The AI can intelligently read and extract data for the following standard attributes:

Document Name & Number
From Date & To Date
Issuing Country & Location
Issued On & Issuing Authority
Issuing Comments

Once the extraction is complete, the fields populate instantly. The user simply reviews the data, makes any necessary adjustments, and submits.

Key Business Benefits

Massive Time Savings: HR teams and employees can cut down data entry time by up to 80% per document.
Improved Data Accuracy: AI extraction eliminates typos, date formatting issues, and transcription errors, ensuring cleaner data in your HCM ecosystem.
Enhanced Data Integrity: By combining this feature with Visual Builder Studio (VBS), you can make the scanned fields read-only after extraction. This ensures that the data in the system perfectly matches the uploaded document, preventing unauthorized manual alterations.

Important Tips & Constraints to Keep in Mind

Before deploying this feature, here are a few technical nuances you should note:

Redwood Only: This feature is exclusive to the Redwood Document Records page and is only available during the creation of a new document record (not while editing an existing one).
Single Attachment Limit: The AI can only scan one attachment per document record at a time. If you upload multiple files to a single record, the “Extract” button will be disabled.
Supported Formats: The AI supports PDF, PNG, JPEG, DOCX, and TXT files. Password-protected documents cannot be scanned.
Flexfield Exclusion: At this time, the extraction logic only maps to standard Oracle attributes; it does not support descriptive flexfields (DFFs) or developer flexfields.
Logic Precedence: If you have configured a business rule to default a value into a specific field, the AI extraction logic will take precedence and overwrite that default value with the data found in the document.

Step-by-Step Guide: How to Enable and Configure

By default, the Enable Extract Data from Attachment property is set to disabled. To turn it on, an administrator will need to use Visual Builder Studio (VBS).

Navigate to the Redwood Document Records page.
Open the Business Rules pane via Visual Builder Studio.
Locate the page property named Enable Extract Data from Attachment.
Toggle the property or set its value to Yes.
Express or publish your changes.

User Experience Flow:

Once enabled, the user experience is incredibly straightforward:

The user selects the Document Type (this triggers the attachment region to become visible).
The user uploads the file.
The Extract button becomes clickable. Clicking it kicks off the AI process and autofills the page.

Security and Access Requirements

To ensure your users have access to this new feature, confirm that they hold the appropriate SaaS roles and aggregate privileges:

User Type	Required Role	Assigned Aggregate Privilege
Employee	Employee	Manage Person Documentation by Worker (`ORA_PER_WORKER_DOCUMENT_MANAGEMENT_DUTY`)
Line Manager	Line Manager	Manage Person Documentation by Manager (`ORA_PER_MANAGE_PERSON_DOCUMENTATION_BY_MANAGER`)
HR Specialist	Human Resource Specialist	Manage Person Documentation (`ORA_PER_DOCUMENT_MANAGEMENT_DUTY`)

Final Thoughts

The 26C update proves once again that Oracle is dedicated to injecting practical, everyday AI into Core HR functions. By eliminating routine data entry from Document Records, HR departments can shift their focus from administrative paperwork to strategic initiatives.

Are you planning to enable AI Document Extraction in your next upgrade cycle? Let me know your thoughts or questions in the comments below!