Purchases - Enhanced AI Invoice Scanning
Overview
The Enhanced AI Invoice Scanning feature improves the speed and accuracy of processing purchase invoices within Efimis.
The AI engine can now:
- Extract invoice data from uploaded PDFs
- Process multi-page invoices
- Improve supplier recognition
- Suggest smarter GL allocations
- Identify matter references
- Improve allocation consistency using historical posting behaviour
This enhancement is particularly useful for:
- High-volume invoice processing
- Multi-line invoices
- Matter-related disbursements
- Repetitive supplier transactions
Key Enhancements
| Feature | Description |
|---|---|
| Multi-Page Invoice Scanning | AI can now process invoices containing multiple pages |
| Improved Matter Identification | Detects matter references and allocation targets from invoice content |
| Enhanced Supplier Matching | Improved recognition of existing suppliers and similar supplier names |
| Smarter GL Allocation Suggestions | Uses AI logic and historical posting behaviour to suggest allocations |
| Improved Tax Code Handling | Prevents valid AI tax code selections from being overwritten |
| Allocation Validation Rules | Prevents invalid GL suggestions that would normally fail manual posting validation |
How AI Allocation Suggestions Work
The AI engine evaluates multiple data sources before suggesting a GL allocation.
Allocation Signals Used
| AI Signal | Description |
|---|---|
| Supplier History | Reviews accounts previously used for the supplier |
| Historical Invoice Lines | Compares similar invoice line descriptions from earlier postings |
| GL Account Matching | Matches invoice wording against GL account names and descriptions |
| Firm-Wide Fallback Accounts | Includes commonly used firm accounts where historical data is limited |
Multi-Page Invoice Support
Supported Functionality
- Single PDF uploads
- Multi-page invoices
- Large line-item invoices
- High-volume supplier invoices
Benefits
- Reduced manual entry
- Improved extraction accuracy
- Better handling of detailed supplier invoices
Matter Allocation Recognition
The AI scanner can identify:
- Matter references
- Matter numbers
- Allocation targets
This improves:
- Matter disbursement coding
- Client-related invoice allocation
- Cost recovery processing
Supplier Recognition Improvements
The enhanced supplier matching process helps:
- Detect existing suppliers more accurately
- Reduce duplicate supplier selection
- Improve historical matching
Important
If a supplier cannot be confidently matched:
- Manual review may still be required
- Supplier creation may require approval depending on firm policy
Tax Code Handling
Enhanced Behaviour
Where AI successfully identifies a line-item tax code:
- The tax code will remain applied
- Selecting a GL account will not automatically overwrite the tax code
This helps improve:
- VAT/GST accuracy
- Invoice coding consistency
- Reduced manual correction
Allocation Validation Rules
The AI scanner follows the same restrictions as manual posting.
The AI Will NOT Suggest
- System control accounts
- Client journal GL codes
- GL codes from unrelated office journals
- GL accounts with a Bank account type
Purpose
This ensures:
- Posting compliance
- Accounting integrity
- Consistent validation rules
Manual Review Still Required
AI assists with invoice processing, but invoices should still be reviewed before posting.
Users Should Validate
- Supplier selection
- Invoice totals
- Matter allocations
- GL allocations
- Tax codes
Best Practices
Recommended
- Upload clear PDF invoices
- Use consistent supplier naming
- Review AI-generated allocations
- Validate matter allocations carefully
- Maintain consistent GL coding practices
Avoid
- Poor quality scans
- Blurry image PDFs
- Incomplete invoice documents