How does the AI handle non-standard accessorial charges?
expand_more
Our NLP models are trained specifically on carrier invoices to recognize common accessorial charges like "re-weight," "lift-gate," "inside delivery," "residential delivery," and "detention fees" regardless of how each carrier names them. The system maintains a comprehensive mapping of over 500 accessorial charge variations across major carriers including FedEx, UPS, DHL, and regional LTL providers. When a charge appears on an invoice that isn't pre-approved in your carrier contract, the system automatically flags it for review and provides a side-by-side comparison with your contracted rates. This proactive approach has helped our clients recover an average of 3.2% of their annual freight spend from incorrect accessorial charges.
What is the typical time from ingestion to ERP posting?
expand_more
Standard invoices are processed end-to-end in under 3 minutes from the moment they hit your inbox. This includes document ingestion, AI-powered data extraction, validation against your business rules, GL coding assignment, and final posting to your ERP. For invoices that trigger audit flags or require manual review, our Human-in-the-Loop (HITL) workbench ensures resolution within your custom SLA - typically under 4 hours for most clients. The workbench is designed for speed with keyboard shortcuts, pre-filled correction suggestions, and one-click approval workflows. Our enterprise clients process over 10,000 invoices monthly with an average time-to-approval of just 2.8 minutes.
Can you detect duplicate invoices across different carriers?
expand_more
Yes, our duplicate detection engine goes far beyond simple invoice number matching. We use advanced fuzzy matching algorithms that analyze multiple data points including shipment IDs, Bill of Lading (BOL) numbers, PRO numbers, pickup/delivery dates, origin-destination pairs, and invoice amounts. This multi-factor approach catches duplicates even when a freight broker and the underlying carrier both submit invoices for the same shipment with completely different invoice numbers. The system also identifies partial duplicates - such as when line items from one invoice appear on another - and presents them side-by-side for quick resolution. On average, our clients prevent 1.5% of duplicate payments annually through this detection capability.
Which ERP systems do you support?
expand_more
We offer pre-built, certified connectors for the most popular enterprise ERP systems including SAP S/4HANA, SAP Business One, Oracle NetSuite, Microsoft Dynamics 365 Business Central, Microsoft Dynamics 365 Finance & Operations, Sage Intacct, QuickBooks Enterprise, and Costar. Each connector supports bi-directional sync for vendor master data, GL codes, cost centers, and payment status updates. For organizations with custom or legacy ERP systems, our REST API provides complete flexibility with comprehensive documentation, SDKs in Python, Node.js, and .NET, plus dedicated integration support. Typical ERP integration takes 2-4 weeks from kickoff to go-live, with our implementation team handling the heavy lifting.
How do you handle multi-currency logistics invoices?
expand_more
Our platform is built for global logistics operations with full multi-currency support. When processing an invoice, the system automatically detects and extracts the original currency from the document - whether it's USD, EUR, GBP, CAD, MXN, or any of 150+ supported currencies. You can configure your preferred exchange rate policy: real-time rates from trusted providers like Reuters and Bloomberg, daily fixed rates, monthly average rates, or custom rates imported from your treasury system. The platform maintains a complete audit trail showing both the original currency amount and the converted functional currency for GL reporting. For organizations with complex intercompany billing, we support multi-entity configurations with different functional currencies per legal entity.
What is your extraction accuracy for blurry scans?
expand_more
Our AI extraction engine uses state-of-the-art vision transformer technology specifically trained on logistics documents - including faded thermal prints, low-resolution faxes, smartphone photos of BOLs, and multi-generation photocopies. Unlike traditional OCR that struggles with imperfect documents, our models achieve 99.2% accuracy on standard invoices and maintain 94%+ accuracy even on severely degraded documents. Each extracted field includes a confidence score, and any field falling below your configured threshold (typically 95%) is automatically highlighted for human review. The system continuously learns from corrections made by your team, improving accuracy over time for your specific carrier formats. We also provide document quality analytics so you can work with carriers to improve their invoice submission quality.
