Intake and classification
A dedicated mailbox, a Drive or SharePoint folder, or an upload form. Each file is classified by type, such as invoice, credit note or delivery note, and matched to its supplier.
Invoices arrive by email, as PDFs or phone photos, and someone types them in. We automate reading and checking, and only send you what doesn't add up.
Invoice and document automation reads PDFs, scans and e-invoices, extracts the fields you need and writes them to your ERP or a spreadsheet. We set it up for companies that type in supplier invoices or other repetitive documents by hand today. You start with a 7-day trial on one document type and one intake channel.
A dedicated mailbox, a Drive or SharePoint folder, or an upload form. Each file is classified by type, such as invoice, credit note or delivery note, and matched to its supplier.
If the invoice carries structured data (Facturae, UBL or Factur-X), we read the XML. If it's a PDF or a photo, we use OCR with Google Document AI, Azure AI Document Intelligence or Amazon Textract, or vision models that return the fields in a fixed schema.
Only what needs it goes to review: uncertain readings, totals that don't add up, new suppliers, large amounts and, above all, changed bank details, a classic fraud signal. The reviewer sees the document beside the extracted fields, corrects and approves it, and everything is logged.
Invoices contain personal data, for example from sole traders. We extract only what's needed, use an EU region where one exists, restrict access and limit how long n8n keeps execution data. This is technical assistance, not legal advice: the legal basis, processor agreements and retention periods are yours to decide with your advisor.
When you regularly receive invoices, delivery notes, orders or other documents with recognizable fields, someone spends hours typing them in, and the destination has an API or a well-structured spreadsheet.
A realistic scope: invoices arriving at one specific mailbox are read, validated and written to a Google Sheet or your ERP sandbox, with their exception queue. We review the results on your sample with you, field by field. Scope, dates and terms are set in writing before we start.
It depends on the quality and variety of your documents, so we don't give a number up front. We measure it on your real sample, field by field, and set thresholds: anything that fails the checks goes to human review instead of into the ERP. E-invoices with XML don't even need OCR.
No. We automate data capture and export, but we don't do bookkeeping or tax advice. Posting and allocating each expense, and deciding how an invoice is treated for tax purposes, stay with your team or your accountant; the system saves them the typing and flags what doesn't add up.
Any that has an API or accepts file imports, such as Odoo, Holded, Sage or Business Central, or simply Google Sheets or Excel. If your software only allows manual entry, we look at file imports before any fragile workaround, and we check what your specific version allows.
In your systems: the original PDF stays in your folder or ERP, and the workflows run on a server you pay for. The OCR service or model we choose processes each document under its own terms, which we review with you, and we limit what's kept in the execution history.
Tell us how they arrive today and where they need to go. We'll send a written 7-day trial proposal with one document type and your real sample.
Pick a goal and we'll suggest the services that fit best.