Invoice Upload & OCR

Learn how to upload invoices and use SOPHIOS’s AI-powered OCR to automatically extract invoice data.

Overview

SOPHIOS uses advanced Optical Character Recognition (OCR) powered by private AI infrastructure to automatically extract data from invoices, reducing manual data entry by up to 80%.

Uploading an Invoice

OCR Data Extraction

What Gets Extracted

SOPHIOS’s AI automatically identifies and extracts:

Vendor Information:

• Vendor name
• Vendor address
• Vendor contact details
• VAT/Tax ID number

Invoice Details:

• Invoice number
• Invoice date
• Due date
• Purchase order number (if present)

Line Items:

• Item descriptions
• Quantities
• Unit prices
• Line totals

Financial Data:

• Subtotal
• Tax amount(s)
• Total amount
• Currency

Additional Data:

• Payment terms
• Bank details (if present)
• Notes and references

Reviewing OCR Results

After processing, review the extracted data:

Categorizing Invoices

Automatic Categorization

SOPHIOS suggests categories based on:

• Vendor Name - Recognizes known suppliers
• Description Keywords - Matches item descriptions
• Historical Patterns - Learns from past categorizations
• Machine Learning - AI-powered suggestions

Manual Categorization

Override auto-suggestion:

1. Click “Category” dropdown
2. Select from available categories
3. Create new category if needed

Common Categories:

• Fuel & Lubricants
• Maintenance & Repairs
• Marina/Port/Hangar Fees
• Catering & Provisions
• Insurance
• Crew Expenses
• Equipment & Supplies

Assigning to Budgets

Link invoices to budgets for tracking:

Invoice Status Workflow

Invoices progress through these statuses:

PENDING → PROCESSING → NEEDS_REVIEW → VERIFIED → APPROVED → EXECUTED

Status Definitions:

Invoice Approval Workflow

For Accountants

1. Review OCR-extracted data
2. Verify against original invoice
3. Correct any errors
4. Change status to VERIFIED
5. Submit for approval

For Approvers (Admin/Owner)

1. Review verified invoices
2. Check budget impact
3. Approve or reject with notes
4. Mark as APPROVED

For Executors (Owner/Admin)

1. Review approved invoices
2. Confirm payment has been made
3. Mark as EXECUTED

Expense Classification

Expense Type

Each invoice can be classified by expense type:

• OPEX - Operational expenditure (day-to-day costs)
• CAPEX - Capital expenditure (investments, improvements)
• MIXED - Invoices containing both OPEX and CAPEX items

Spending Line Assignment

Invoices can be assigned to spending lines (cost centers) for granular cost tracking:

• Assign during invoice creation or editing
• Change spending line assignment at any time via the inline badge
• View spending breakdowns in the Financial Reports dashboard

Advanced Features

Duplicate Detection

SOPHIOS automatically detects potential duplicates:

• Same vendor + invoice number
• Same vendor + similar amount + similar date
• Warns before saving

Anomaly Detection

AI flags unusual transactions:

• Amount significantly higher than usual
• New vendor (never seen before)
• Missing required fields
• Unusual payment terms

Vendor Matching

Automatic vendor management:

• Creates vendor records from invoices
• Links all invoices to vendors
• Tracks spending per vendor
• Suggests vendor name corrections

OCR Accuracy Tips

Best Practices for Scanning

To maximize OCR accuracy:

1. File Format: Use PDF when possible
2. Resolution: Scan at 300 DPI or higher
3. Orientation: Ensure invoice is right-side up
4. Lighting: Good lighting for photos, no shadows
5. Focus: Clear, not blurry images
6. Cropping: Include entire invoice, minimal margins
7. Color: Color or grayscale (better than black & white)

What Reduces Accuracy

Avoid these common issues:

• ❌ Blurry or out-of-focus images
• ❌ Poor lighting or shadows
• ❌ Handwritten invoices (not supported)
• ❌ Skewed or rotated images
• ❌ Low resolution (under 150 DPI)
• ❌ Watermarks obscuring text
• ❌ Crumpled or damaged invoices

Handling Different Invoice Formats

Structured Invoices

Best OCR Performance:

• Standard invoice templates
• Clear table structures
• Printed text
• Professional formats

Examples:

• Fuel company invoices
• Marina/port invoices
• Supplier invoices from major vendors

Unstructured Invoices

More Manual Review:

• Non-standard formats
• Receipts without headers
• Invoices in unusual layouts

Best Practice: For unstructured invoices, verify all fields carefully.

Multi-Page Invoices

SOPHIOS handles multi-page PDFs:

• Processes all pages automatically
• Extracts items from multiple pages
• Combines into single invoice record

Editing After Save

Edit Invoice Data

After saving, you can still edit:

1. Navigate to Invoices list
2. Click the invoice to open details
3. Click “Edit” button
4. Modify any field
5. Click “Save Changes”

Re-Process OCR

If OCR failed or produced poor results:

1. Open invoice details
2. Click three-dot menu (⋮)
3. Select “Re-process OCR”
4. Upload new/better quality file (optional)
5. Review new results

Bulk Upload

For multiple invoices:

Invoice Attachments

Attach Supporting Documents

Add additional files to invoices:

• Purchase orders
• Delivery notes
• Contracts
• Payment confirmations

How to attach:

1. Open invoice details
2. Scroll to “Attachments” section
3. Click “Add Attachment”
4. Upload file(s)

Exporting Invoices

Export Options

Single Invoice:

1. Open invoice details
2. Click three-dot menu (⋮)
3. Select “Export as PDF” or “Download Original”

Multiple Invoices:

1. Go to Invoices list
2. Select checkboxes for invoices
3. Click “Export Selected”
4. Choose format (PDF, CSV, Excel)

Troubleshooting OCR

OCR Failed

If OCR fails completely:

1. Check file format - Must be PDF, JPG, or PNG
2. Check file size - Must be under 10 MB
3. Check image quality - Must be clear and legible
4. Try re-upload - Sometimes temporary glitch
5. Manual entry - Enter data manually as fallback

Low Accuracy

If OCR accuracy is poor:

1. Check original quality - Rescan if blurry
2. Convert to PDF - Photos work best as PDFs
3. Crop properly - Remove excess whitespace
4. Rotate correctly - Ensure upright orientation
5. Report issue - Contact support with sample

Missing Fields

If OCR missed some fields:

1. Check original - Field might genuinely be missing
2. Manual entry - Type in missing data
3. Mark incomplete - Flag for later completion

Common Questions

Does OCR work in multiple languages?

Yes, SOPHIOS OCR supports:

• English
• Spanish
• French
• Italian
• German
• Portuguese
• (More languages being added)

Can I edit OCR results?

Absolutely! Always verify and correct OCR results. Your corrections train the AI.

What if my invoice has no text?

Image-only invoices with very poor quality may have limited OCR support. Manual entry recommended for handwritten documents.

How secure is invoice upload?

Very secure:

• Encrypted transfer (HTTPS/TLS)
• Encrypted storage (AES-256)
• Access control (role-based permissions)
• Audit trail (all actions logged)

Next Steps:

• Create a budget
• Set up recurring expenses
• View financial reports