Sicada's Document Intelligence platform parses structured, semi-structured, and unstructured documents across financial, legal, identity, operational, and compliance categories. The core document types supported out of the box cover the highest-volume use cases across enterprise operations with custom model support available for document types outside the standard library.
The full supported list, field extraction detail, and format requirements for your specific use case are available from the Sicada team on request.
Supported Document Categories
Financial Documents
Financial documents are the highest-volume document intelligence use case across enterprise operations and the category with the strongest ROI case for automation, given that manual invoice processing costs between $15 and $40 per document.
- Invoices- Vendor name, invoice number, line items, unit prices, tax amounts, totals, due dates, payment terms, and purchase order references. Sicada parses invoices to extract structured data from documents including multi-page, multi-vendor, and non-standard format invoices that template-based systems cannot handle. Extraction accuracy on structured invoices reaches 99% on clean digital documents; scanned and handwritten invoices are processed with confidence scoring to flag low-accuracy fields for human review before downstream payment processing.
- Purchase Orders- PO number, vendor details, line item descriptions, quantities, agreed prices, delivery terms, and approving authority. Cross-document validation against corresponding invoices is supported- flagging mismatches between ordered and invoiced quantities or amounts before they reach accounts payable.
- Bank Statements- Account number, statement period, opening and closing balances, transaction-level line items including date, description, debit, and credit. Extracted for financial reconciliation, credit assessment, and fraud detection workflows.
- Receipts- Merchant name, transaction date, itemised line items, subtotal, tax, and total. Extracted for expense management, reimbursement workflows, and audit trail generation.
- Pay Slips- Employee name, employee ID, pay period, gross pay, deductions, net pay, and employer details. Processed for payroll reconciliation, lending affordability assessments, and HR record management.
Legal and Contractual Documents
To parse contracts AI requires extraction capability well beyond simple text recognition- legal documents carry dense, variable-structure content where the same clause can appear in dozens of structural configurations across different document templates.
- Contracts and Agreements- Party names, effective dates, term length, payment clauses, termination conditions, liability caps, jurisdiction, governing law, and non-standard clause flags. Sicada's contract parsing identifies and extracts key commercial and risk terms, surfaces clauses that deviate from standard templates, and routes documents requiring legal review based on extracted clause content rather than requiring manual triage.
- NDAs and Confidentiality Agreements- Disclosing and receiving party details, scope of confidential information, duration of confidentiality obligation, permitted disclosures, and breach remedies. Commonly processed as part of vendor onboarding and partnership workflows.
- Lease and Tenancy Agreements- Landlord and tenant details, property address, lease term, monthly rent, deposit amount, renewal conditions, and break clauses. Extracted for real estate portfolio management and legal compliance tracking.
- Terms and Conditions- Liability clauses, indemnification terms, dispute resolution mechanisms, and amendment provisions. Flagged against a defined acceptable terms baseline for procurement and vendor management workflows.
Identity and KYC Documents
Identity document parsing is a foundational requirement for KYC, AML compliance, and customer onboarding workflows across financial services, insurance, healthcare, and legal sectors.
- National Identity Cards- Full name, date of birth, ID number, nationality, issue date, expiry date, and MRZ data where present. Supported for Indian Aadhaar, Emirates ID, Saudi National ID, Singaporean NRIC, and standard EU identity card formats.
- Passports- Name, nationality, passport number, date of birth, date of issue, expiry date, and full MRZ extraction for automated validity verification.
- Driver's Licences- Name, address, licence number, vehicle categories, issue and expiry dates. Supported across major international formats including UK, US, UAE, Indian, and Australian driver's licence layouts.
- Visas- Visa type, issue and expiry dates, permitted stays, and entry conditions. Extracted for employee onboarding, immigration compliance, and travel management workflows.
Insurance and Claims Documents
Insurance is one of the document intelligence supported documents categories with the highest volume and the strongest operational case for automation. Manual claims processing costs between $10 and $25 per document and is subject to high rates of data entry error.
- Claim Forms- Claimant details, policy number, incident date, incident description, claimed amounts, and supporting documentation references. Extracted for claims triage, automated routing, and fraud scoring.
- Policy Documents- Policy number, insured name, coverage type, coverage limits, premium, effective and expiry dates, and exclusions. Parsed for renewal management, portfolio analytics, and compliance reporting.
- Medical Reports and Discharge Summaries- Patient name, diagnosis codes, treatment details, physician name, and admission and discharge dates. Processed for health insurance claims validation and clinical documentation workflows.
- Damage Assessment Reports- Assessor details, assessment date, damage description, repair cost estimates, and photographic evidence references. Extracted for property and automotive claims settlement workflows.
HR and Onboarding Documents
- Employment Contracts- Employee name, role title, start date, salary, working hours, probation period, and termination terms.
- Offer Letters- Candidate name, role, compensation package, start date, and reporting line.
- Educational Certificates and Transcripts- Institution name, candidate name, qualification title, grade or classification, and issue date. Extracted for credential verification in hiring workflows.
- Reference Letters- Referee name, relationship to candidate, dates of employment, and key competency references.
Healthcare Documents
- Prescription Forms- Patient name, prescribing physician, medication names, dosages, frequency, and issue date. Processed for pharmacy management and insurance claim validation.
- Consent Forms- Patient name, procedure description, consenting physician, date, and patient signature status. Extracted for clinical governance and compliance workflows.
- Lab Reports- Test names, results, reference ranges, ordering physician, and report date. Parsed for clinical data integration and insurance claims processing.
Government and Compliance Documents
- Tax Documents- Tax identification numbers, filing periods, declared income, deductions, and assessable amounts. Supported formats include GST invoices for Indian compliance, VAT returns for Gulf markets, and standard corporate tax filing documents.
- Regulatory Filings- Company registration documents, annual return filings, and compliance certificates extracted for corporate governance and vendor due diligence workflows.
- Tender Documents and RFPs- Scope of work, submission deadlines, evaluation criteria, and pricing schedules extracted for procurement management.
File Formats Supported
Sicada's Document Intelligence ingests documents in the following formats:
- PDF- both digital-native and scanned, including multi-page documents
- TIFF and JPEG- scanned image formats including multi-page TIFF
- PNG- image captures of physical documents
- DOCX and DOC- Microsoft Word formats for contract and correspondence processing
- XLSX and XLS- spreadsheet formats for financial and operational data documents
- PPTX- presentation formats for proposal and tender document processing
For scanned or photographed documents, OCR preprocessing is applied before extraction. Documents with degraded scan quality- low resolution, skewed angles, heavy background noise are processed with confidence scoring at the field level, with low-confidence fields flagged for human review rather than silently passed downstream.
What If Your Document Type Is Not Listed?
The document categories above cover the standard library of pre-trained extraction models. For document types outside this library- proprietary forms, industry-specific templates, custom internal documents- Sicada supports custom model training.
Custom model training requires a minimum sample set of the target document type. The model learns the structural and content patterns of the specific document format and produces field-level extraction configured to your exact data requirements. Custom models are typically production-ready within five to ten business days of sample submission, depending on document complexity and field count.
Summary: What Sicada Document Intelligence Parses
- Financial: Invoices, purchase orders, bank statements, receipts, pay slips
- Legal: Contracts, NDAs, lease agreements, terms and conditions
- Identity and KYC: Passports, national IDs, driver's licences, visas
- Insurance and claims: Claim forms, policy documents, medical reports, damage assessments
- HR and onboarding: Employment contracts, offer letters, certificates, reference letters
- Healthcare: Prescriptions, consent forms, lab reports
- Government and compliance: Tax documents, regulatory filings, tender documents
- File formats: PDF, TIFF, JPEG, PNG, DOCX, XLSX, PPTX
Frequently Asked Questions
What document types does Sicada Document Intelligence support?
Sicada's Document Intelligence platform supports financial documents including invoices, purchase orders, bank statements, and receipts; legal documents including contracts and NDAs; identity documents including passports, national IDs, and driver's licences; insurance documents including claim forms and policy documents; HR documents including employment contracts and certificates; healthcare documents including prescriptions and lab reports; and government compliance documents including tax filings and regulatory submissions.
Can Sicada parse contracts AI-style to extract specific clauses?
Yes. Sicada's contract parsing extracts key commercial and risk terms- party names, effective dates, payment clauses, termination conditions, liability caps, jurisdiction, and non-standard clause flags and routes documents requiring legal review based on extracted content rather than requiring manual triage.
How does Sicada extract structured data from invoices?
To parse invoices AI extraction covers vendor name, invoice number, line items, unit prices, tax amounts, totals, due dates, payment terms, and PO references. Multi-page and non-standard format invoices are supported. Confidence scoring at the field level flags low-accuracy extractions for human review before they enter downstream payment workflows.
What happens if my document type is not in the standard library?
Sicada supports custom model training for document types outside the standard library. A minimum sample set of the target document type is required. Custom models are typically production-ready within five to ten business days of sample submission depending on document complexity.
