Document Extraction & Structured Data

AI-Driven · PII Redaction · Batch Extraction

From Mixed Documents to
Structured, Compliant Data

Batch-import mixed-format documents, including PDF, Word, and PPT files,
and let AI extract key data, redact PII, and generate structured Excel
and sanitized PDF files.

100

Documents Processed

Field Types Extracted

PII Instances Masked

From Manual Review to
Automated Document Processing

The AI agent automates batch upload, extraction, PII redaction,
and structured output in one workflow.

The manual route today

1 Open each file, read every page hunting for names, IDs and addresses
2 Black out or delete fields one by one, hoping nothing is missed
3 Re-key the extracted fields into spreadsheets by hand
4 Keep manual logs of what was cleaned — for the auditors

Typical cost for 100 documents: hours of manual review and a real risk of missed PII.

With the agent

1 Drop in a batch folder — 100 mixed PDF / Word / PowerPoint files
2 The agent extracts fields and flags 89 PII items across 20+ categories
3 Redaction is applied in place — masked or erased at the byte level
4 A human approves the checkpoint; the Excel list and PDF archive are delivered

With the agent: ~3 minutes per 100 documents, with an audit trail for every redaction

Run the Demo

Upload your documents and experience automated
extraction and PII redaction with the Agent.

Employee
Resume.pdf

Emergency
Contacts.docx

Employee
Information Form.xlsx

Start

Bank & Payroll
Information.pdf

Professional
Certifications.docx

Position Appointment
Materials.pptx

Parsing the document (1/4)

Parse external PDF contracts / Parse internal Word templates

Process completed !

Result Preview

Structured Checklist.xlsx

Employee Record_Redacted Archive.pdf

Employee Resume_Redacted Archive.pdf

Note: The sample documents on the page are used solely to demonstrate product features,
and all related content does not correspond to real entities or business information.

Four Steps to Automated
Document Processing

Explore the Agent pipeline in workflow view.

Parse the Batch

Upload 100 PDF, DOCX, and PPT files at once, with native parsing for tables, layouts, and scanned pages.

Classify Fields

Extract names, emails, phones, degrees, and certifications into a fixed schema.

Redact & Review

Detect 20+ PII types, automatically redact sensitive data, and flag high-risk cases for human review.

Deliver Files

Generate a structured Excel workbook and redacted PDFs, with a complete audit trail.

Explore Agent
Developer Capabilities

An extensible redaction Agent that seamlessly
integrates with existing compliance systems.

Parse

Multi-Format Reading

Batch-read PDF, Word, and PowerPoint files with OCR support.

Understand

Field Classification

Extract target fields using a schema-constrained approach.

Reason

Redaction Decisions

Rules + LLM determine what to mask or remove.

Act

Tool Calling

Function calling triggers redaction and native document generation.

Generate

Native Outputs

Generate native Excel workbooks and redacted PDFs.

// Create a batch processing agent, extract fields, and output desensitized results
string instruction =
"Read PDF, Word, and PPT format documents in batches from attachments, automatically identify and extract key information from each document (such as 'name', 'education', 'qualifications', 'contact information', etc.)," +
"Generate a structured Excel list with one row per document, and desensitize the 'contact information' at the same time. Finally, save and output in both Excel and PDF formats";

// Create AIOptions configuration
AIOptions options = new AIOptions();

// Set SpireToken Key
options.SpireToken = "**************************";

options.WorkDir= "out";

// Using the Workbook object
using (Workbook xls = new Workbook())
{
    // Create an AI document processor
    AIDocumentProcessor processor = xls.AI(options);

    // Execute AI instructions
    processor.ExecuteInstruction(xls, instruction, null, new string[] { "in.pdf", "in.docx", "in.pptx" });
}

Built for Bulk Extraction,
Built for Compliance

an agent built for document-heavy workflows,
with every step in view

Concurrent Multi-Format Parsing

Process PDF, DOCX and PPTX in one batch. Native parsing preserves tables, layout and scanned content — no OCR transcription loss on the fields you need.

Agent-Driven PII Detection

Rules and LLM reasoning detect 20+ PII types, including emails, phone numbers, addresses, and ID numbers, with high-risk cases flagged for review.

Structured Data Output

Extract every record into a fixed schema—Name, Email, Phone, Degree, and Certifications—and export to a native Excel workbook.

Irreversible Redaction

Remove sensitive values at the content layer in native PDFs, rather than simply hiding them behind overlays.

Batch Concurrency

Process up to 100 documents in parallel, with a typical batch completed in about 3 minutes.

Standardized Archiving

Generate redacted PDFs with the original layout, ready for standardized naming and archival.

Common Questions

Which document formats does extraction support?

PDF (including scanned pages, which trigger OCR automatically), Word (.doc/.docx), PowerPoint (.ppt/.pptx), Excel (.xls/.xlsx) and image formats (JPG/PNG/TIFF). Mixed batches are fine — one run can hold 100 files across formats.

How accurate is PII redaction?

The agent recognizes 20+ PII types — emails, phone numbers, ID numbers, addresses, bank accounts and more — using a rules-plus-LLM hybrid. High-risk cases keep a human approval step, so every batch meets compliance-audit expectations.

Can redacted data be recovered?

No. Values are erased irreversibly at the content layer of native PDFs — not hidden or overlain. This satisfies data-minimization and "right to be forgotten" requirements under GDPR and similar privacy laws.

What is the maximum batch size?

Standard runs accept 100 documents per batch, processed in parallel across PDF, Word and PPT. A full 100-file batch typically completes in about 3 minutes.

Can I define my own extraction fields?

Yes. Presets cover resumes, certificates and employee records out of the box, and you can define custom fields, mapping rules and redaction policies through the visual configuration.

How is document data kept secure?

All transfers use TLS 1.3 and storage uses AES-256. Original files can be auto-deleted after processing, and on-premises deployment keeps document data entirely inside your network.

Turn Documents into
Compliant Data Now

Get Your Free 1-Month Developer License

Run it Now Book a demo