Overview
This skill extracts key information from various document types.
When to Use
- •Processing invoices or receipts
- •Extracting contact information
- •Reading form submissions
- •Parsing structured data from documents
Extraction Types
- •Financial Data: Amounts, dates, account numbers
- •Contact Info: Names, emails, phone numbers, addresses
- •Dates & Times: Appointments, deadlines, schedules
- •Identifiers: Order numbers, invoice IDs, reference codes
Output Format
Return extracted data as structured JSON:
json
{
"type": "invoice|contact|form|other",
"extracted_data": {
"amount": 1250.00,
"date": "2026-02-01",
"client": "ABC Corp",
"reference": "INV-2026-001"
},
"confidence": 0.95
}
Validation
- •Cross-reference extracted data with known patterns
- •Flag unusual values for human review
- •Verify data consistency across fields
Error Handling
- •Log extraction failures
- •Note confidence levels for uncertain extractions
- •Fallback to manual review for low-confidence data