Skip to main content
Rivet
Features
Pricing
Support
About
Join Waitlist
RIVET Support

Documentation & Help Center

Contact
Support Home

Quick Links

FAQGlossary

Documentation

Document Management OverviewUploading DocumentsUnderstanding OCR ProcessingAIA Document ClassificationPosting Documents to Accounting
Support Home
Documents & OCR
Understanding OCR Processing
Documents & OCR

Understanding OCR Processing

How Mindee OCR extracts data from your documents automatically

6 minLast updated: 2026-01-16
ocrmindeeextractionconfidenceprocessingdata

RIVET uses Mindee's advanced OCR technology to automatically extract data from your documents. This guide explains how OCR processing works, what data is extracted, and how to review and correct the results.

How OCR Works

When you upload a document, RIVET sends it to Mindee's OCR engine for processing. The system analyzes the document layout, identifies text regions, and extracts structured data automatically.

1

Document Analysis

The system identifies document type (invoice, receipt, AIA form, etc.)

2

Text Recognition

OCR reads all text content from the document

3

Data Extraction

Key fields like vendor, amount, and date are identified and extracted

4

Confidence Scoring

Each extracted field receives a confidence score

What Data is Extracted

Depending on document type, RIVET extracts different fields:

FieldDescriptionDocument Types
Vendor NameCompany or supplier nameInvoices, Bills
Invoice NumberUnique document identifierInvoices, Bills
Invoice DateDate document was issuedAll documents
Due DatePayment due dateInvoices, Bills
Total AmountTotal amount dueAll documents
Line ItemsIndividual line descriptions and amountsInvoices, AIA forms
Tax AmountSales tax or other taxesInvoices

Understanding Confidence Scores

Each extracted field has a confidence score from 0% to 100% indicating how certain the OCR engine is about the extraction.

90-100% - High confidence, likely accurate
70-89% - Medium confidence, verify before use
Below 70% - Low confidence, manual review required

Always Verify Low Confidence

Fields with confidence below 70% are highlighted in the review screen. Always verify these values against the original document before posting to accounting.

Reviewing Extracted Data

To review and correct extracted data:

  1. Navigate to Documents and click on a processed document
  2. Review the extracted fields displayed alongside the document preview
  3. Click on any field to edit the value if incorrect
  4. Fields with low confidence are highlighted for your attention
  5. Click Save to update the extracted data

Making Manual Corrections

If OCR extraction is incorrect or incomplete, you can manually edit any field:

Edit Fields

Click on any extracted field to edit its value. Your changes are saved and will be used when posting to accounting.

Add Missing Data

If a field wasn't extracted, you can manually enter the value. This is common for handwritten information.

Improve Future Extractions

If you consistently need to correct the same type of error, consider improving your scan quality or using a different document format.

Need help? Contact our support team at support@slpwlk.io

Related Articles

documentsDocument Management OverviewIntroduction to RIVET's document management and OCR capabilitiesRead moredocumentsUploading DocumentsHow to upload invoices, receipts, and AIA documents for processingRead moredocumentsAIA Document ClassificationHow RIVET automatically classifies G702, G703, and change order documentsRead more
Can't find what you're looking for?
Contact Support

Product

  • Features
  • Pricing
  • Integrations

Solutions

  • For Electricians
  • For Plumbers
  • For HVAC
  • For General Trades

Company

  • About
  • Contact
  • Security

Resources

  • Support Center
  • Templates
  • Blog

Compare

  • vs Knowify
  • vs Werx

Legal

  • Privacy Policy
  • Terms of Service
Rivet

© 2026 Rivet, Inc. All rights reserved.