Solutions
Invoice & Receipt Processing
Manual invoice processing doesn't scale. Skilled workers are expensive and scarce. Error rates increase with volume.
The Solution
feld.ai reads invoices, extracts fields (supplier, amount, line items, IBAN), matches against master data, and delivers posting suggestions.
How It Works
1. Ingest documents — Import invoices, receipts and credit notes via upload, email or scan folder. feld.ai processes PDF, image and electronic formats (ZUGFeRD, XRechnung).
2. Extract fields — AI recognises supplier, invoice number, date, line items, amounts and IBAN — even on unseen layouts without prior templates.
3. Match master data — Extracted data is automatically checked and enriched against your supplier, account and cost-centre master data.
4. Validate and export — Results are validated via human-in-the-loop UX. Corrections feed back into the model. Export as posting suggestions to your ERP or accounting system.
What sets us apart
Your Document Types
Works with your document types — not just standard invoices
Master Data Matching
Mapping to existing data (supplier IDs, accounts, cost centers)
Human-in-the-Loop
Correction UX for fast validation
EU Hosting
No US cloud provider. Hosting in Austria. ISO 27001 in certification.
Related Solutions
Document Classification — Automatically identify and sort incoming documents.
Data Extraction from Files — Extract structured data from any document type.
Industry Solution: Logistics & Freight — Process freight invoices, CMR notes, and customs documents.
Frequently Asked Questions
Which document types are supported?
feld.ai processes incoming invoices, credit notes, delivery notes, reminders and payment advices. Besides PDF and image files, electronic formats such as ZUGFeRD and XRechnung are natively supported. New document types can be trained within days.
What is the recognition rate?
In production environments, feld.ai achieves a header field recognition rate of over 96%. Line-item extraction reaches over 90% correctly posted lines. Recognition improves continuously through human-in-the-loop corrections.
Is there a DATEV integration?
feld.ai generates posting suggestions in DATEV-compatible format. Export can go directly to DATEV Unternehmen online or via your existing middleware. A native DATEV interface is in preparation.
How long does a pilot project take?
A typical pilot starts with your real documents and runs four to six weeks. In the first week we train the model on your document types, followed by a productive test phase with ongoing optimisation.
Where is data processed?
All data is processed exclusively on our own hardware in Austria. No US cloud provider is involved. ISO 27001 certification is in preparation.