Intelligent Document Processing
Intelligent Document Processing
Available from: v1.0.64
The Intelligent Document Processing (IDP) feature automates the ingestion, classification, data extraction, and filing of documents across HR, finance, and operations workflows. It combines Optical Character Recognition (OCR) with Natural Language Processing (NLP) to deliver high-accuracy, low-touch document handling at scale.
Overview
Manually processing documents — invoices, contracts, HR forms, receipts, and more — is time-consuming and error-prone. Starting in v1.0.64, the platform handles this automatically:
- Documents are ingested via upload, email integration, or connected storage.
- OCR extracts raw text from scanned images, PDFs, and other document formats.
- NLP classifies the document into the appropriate category (e.g. invoice, employment contract, expense report).
- Key data fields are extracted and mapped to the relevant platform records.
- The document is filed automatically in the correct location within the system.
Key Capabilities
Automatic Document Classification
Documents are analysed upon ingestion and assigned a document type without any manual tagging. The NLP model considers document structure, language, and contextual signals to determine the correct classification.
Supported document categories include:
- Invoices and purchase orders
- Employment contracts and offer letters
- HR forms (onboarding, leave requests, appraisals)
- Expense reports and receipts
- Compliance and regulatory documents
Data Extraction
Once classified, structured data is extracted from each document automatically. Extracted fields are validated and surfaced for review before being committed to platform records.
Examples of extracted data:
- Invoice: vendor name, invoice number, line items, amounts, due date
- Contract: parties, effective date, termination clauses, key obligations
- HR form: employee name, department, manager, dates, approval status
Automated Filing
Processed documents are routed to the correct module and record within the platform — for example, an invoice is linked to the corresponding vendor and payment workflow, while a signed contract is attached to the relevant employee or supplier record.
Benefits
| Benefit | Detail |
|---|---|
| Reduced manual effort | Routine document processing tasks are fully automated |
| Higher accuracy | OCR + NLP pipelines reduce transcription and classification errors |
| Faster processing | Documents are handled in near real-time upon ingestion |
| Audit trail | Every automated action is logged for compliance and review |
| Cross-functional | Works across HR, finance, and operations document types |
How It Works
Document Ingested
│
▼
OCR Engine
(text extraction)
│
▼
NLP Pipeline
(classification + field extraction)
│
▼
Validation Layer
(confidence scoring + exception flagging)
│
┌─┴─────────────┐
▼ ▼
Auto-filed Flagged for
to record manual review
Documents that fall below the confidence threshold are flagged for human review rather than automatically filed, ensuring that edge cases are handled safely.
Exception Handling
Not every document will be processed with full confidence. The platform handles low-confidence cases as follows:
- Documents below the classification confidence threshold are placed in a Review Queue.
- Operators are notified and can confirm, correct, or reject the suggested classification and extracted data.
- Reviewed documents are used to improve future processing accuracy.
Getting Started
- Navigate to Settings → Document Processing to review and configure supported document types.
- Upload a document manually or connect an ingestion source (email, cloud storage).
- Monitor the processing status in the Document Inbox view.
- Review any items in the Review Queue that require manual confirmation.
- Confirmed documents are filed and linked to the relevant platform records automatically.