AI Document Classifier for Odoo: Auto-Tag & Route Files

Document Chaos in ERP

Companies upload thousands of documents to Odoo — invoices, contracts, receipts, specifications, certificates. Without proper classification, finding the right document becomes a needle-in-a-haystack problem. AI classification automatically organizes documents as they arrive.

Classification Categories

# AI document classification types:
# Financial:
#   - Invoice (vendor/customer)
#   - Receipt/expense
#   - Bank statement
#   - Tax certificate
#   - Financial report
# Legal:
#   - Contract/agreement
#   - NDA
#   - Terms of service
#   - Power of attorney
# Operations:
#   - Purchase order
#   - Delivery note
#   - Packing slip
#   - Quality certificate
# HR:
#   - Resume/CV
#   - ID document
#   - Employment contract
#   - Performance review
# Technical:
#   - Product specification
#   - Technical drawing
#   - User manual
#   - Safety data sheet (SDS)

How Classification Works

# AI Document Classification Pipeline:
# 1. Document uploaded to Odoo (Documents module)
# 2. Extract text:
#    - PDF: text extraction or OCR
#    - Image: OCR (Tesseract/Cloud Vision)
#    - Office docs: python-docx/openpyxl
# 3. AI analyzes content:
#    - Document type classification
#    - Entity extraction (dates, amounts, names)
#    - Language detection
# 4. Apply classification:
#    - Set document tag
#    - Move to correct workspace/folder
#    - Link to Odoo record (partner, project, etc.)
# 5. Notify relevant team

Odoo Integration

# Automated action on document creation:
def classify_document(document):
    # Extract text from attachment
    text = extract_text(document.attachment_id)

    # AI classification
    result = ai_classify_document(text)

    # Apply classification
    tag = env['documents.tag'].search([
        ('name', '=', result['document_type'])
    ], limit=1)

    folder = env['documents.folder'].search([
        ('name', '=', result['folder'])
    ], limit=1)

    document.write({
        'tag_ids': [(4, tag.id)] if tag else [],
        'folder_id': folder.id if folder else document.folder_id.id,
    })

    # Extract and link metadata
    if result.get('partner_name'):
        partner = env['res.partner'].search([
            ('name', 'ilike', result['partner_name'])
        ], limit=1)
        if partner:
            document.partner_id = partner.id

    # Extract key data
    if result['document_type'] == 'invoice':
        document.write({
            'res_model': 'account.move',
            'description': f"Amount: {result.get('amount')}"
                          f"\nDate: {result.get('date')}"
                          f"\nVendor: {result.get('vendor')}",
        })

Metadata Extraction

# AI extracts structured data:
# Invoice → amount, date, vendor, invoice number
# Contract → parties, start/end date, value
# Receipt → amount, date, merchant, category
# Resume → name, skills, experience, education
# PO → order number, items, quantities, supplier

# Extracted data stored as document metadata
# Enables searching by extracted fields
# Auto-links to matching Odoo records

Bulk Classification

# Classify existing untagged documents:
def bulk_classify():
    untagged = env['documents.document'].search([
        ('tag_ids', '=', False),
        ('attachment_id', '!=', False),
    ])

    for doc in untagged:
        try:
            classify_document(doc)
        except Exception as e:
            doc.write({
                'description': f'Classification failed: {e}'
            })

    return f'Classified {len(untagged)} documents'

DeployMonkey AI

DeployMonkey's AI agent can set up document classification workflows in Odoo. Automatically organize your document workspace with zero manual tagging.