What Is PEDIF? PDF-to-EDI & E-Invoice Automation | PEDIF

PEDIF explained in more detail

PEDIF closes the gap between documents that arrive or leave as PDFs and systems that need structured data. The following sections describe what PEDIF does, where it fits into an existing system landscape and how teams work with it day to day.

What PEDIF is – and what it is not

PEDIF is a service that turns recurring business documents into structured, validated data and hands that data to the systems a company already uses. It focuses on documents that keep arriving in the same layouts, such as purchase orders from regular customers or invoices from regular suppliers.

PEDIF is not a replacement for an ERP system, an EDI network or a document management system. It complements them where documents still arrive or leave as PDFs. It is also not a general-purpose OCR tool that guesses at every document it sees: known layouts are processed according to defined rules, and anything outside that scope is made visible instead of being passed on silently.

Which documents PEDIF processes

PEDIF works in both directions of a business relationship. On the incoming side it structures documents from customers and suppliers. On the outgoing side it turns a company's own PDF documents into the electronic formats that recipients require.

• Incoming: purchase orders, order changes, order confirmations, delivery notes and supplier invoices
• Outgoing: invoices and credit notes that need to become XRechnung, ZUGFeRD or Peppol e-invoices
• Reverse direction: structured files such as XRechnung, EDIFACT or XML turned into readable PDFs for people
• Further document types, from shipping documents to payment advices, each assessed with its own scope

Where the data goes

The result of processing is not a scanned text file but a structured data set in the format the target system expects. The target format is agreed during onboarding and can differ by document type or recipient.

• Import formats of ERP systems and industry software
• EDI messages such as EDIFACT for partners who work with EDI
• XML, CSV or JSON files for further processing
• REST API and webhooks for direct system-to-system handoff
• E-invoice formats such as XRechnung, ZUGFeRD and Peppol BIS Billing 3.0

How a fingerprint differs from a template or a trained model

A classic template describes fixed positions on a page and breaks as soon as something moves. A trained AI model estimates where information probably is and returns a confidence score, which means the same document can produce slightly different results.

A PEDIF fingerprint describes how a specific recurring layout is recognized and which fields, rules and target output belong to it. Once a fingerprint has been set up and approved, the same input leads to the same output. Information that is not on the document itself, such as an internal customer number, can be added through rules or master data. This makes the result reproducible and traceable, which matters for audits and for automated handoff to other systems.

What happens when a layout changes

Business partners change their documents from time to time: a new logo, an extra column, a different position for the order number. PEDIF does not hide such changes. A document that no longer matches its approved fingerprint is flagged instead of being processed with outdated rules.

The fingerprint is then adjusted, tested with the new documents and approved again. Because fingerprints, validation rules and target outputs are versioned, it stays clear which rules were applied to which document at any point in time.

Example: an incoming purchase order from PDF to ERP

A customer sends a purchase order as a PDF attachment, as it has done for years. The layout is known, so PEDIF recognizes it through the approved fingerprint.

Order number, delivery address, requested delivery date and every order line with article, quantity and price are captured into the agreed fields. The customer's own article numbers are translated into your internal article numbers using master data. Validation checks that mandatory fields are present and that quantities and prices are plausible.

The order is then handed to your ERP system in its import format, where it appears like any other order. If something does not fit, for example an unknown article number, the order is shown for review with the specific finding instead of being imported with an error. The customer does not have to change anything, and your order desk no longer types the order in.

Example: an outgoing invoice from PDF to e-invoice

Your billing system produces invoices as PDFs, and a growing number of customers require e-invoices. PEDIF takes the PDF invoice as it is, recognizes the approved layout and structures invoice number, dates, parties, lines, taxes and totals.

Depending on the recipient, PEDIF creates an XRechnung, a ZUGFeRD file or a Peppol invoice, validates it against the rules of the target format and checks that totals and tax amounts reconcile. Valid invoices are delivered through the agreed channel; invoices with findings are held back for review. Your billing system and your invoice layout stay unchanged.

What PEDIF needs from you to get started

• Sample documents for each document type and recurring layout in scope, including typical special cases
• A description of the target format your system expects, or an example import file
• Master data needed for enrichment, such as customer or article number mappings, where applicable
• A contact person who reviews exceptions and approves the results during testing

Where people stay involved

Automation does not mean that nobody looks at the process any more. It means that people only need to look where their judgment is actually required.

• Validation findings: missing mandatory data, totals that do not reconcile or values that conflict with master data
• New layouts that do not have an approved fingerprint yet
• Variable inputs such as order text in an email body, spreadsheets or Word files, handled through pedif+ with focused review
• Business decisions, for example whether an unusual order should be accepted

Who works with PEDIF

• Finance and accounts payable teams that want supplier invoices as structured data
• Order desk and sales operations teams that receive customer orders as PDFs
• IT, ERP and EDI teams that need to connect partners who do not use EDI
• Software vendors that want to offer PDF automation inside their own product

Frequently asked questions about PEDIF

Do we need to change our ERP system to use PEDIF?

No. PEDIF delivers data in the format your existing ERP, EDI or DMS processes already expect. The target format is agreed during onboarding.

Do our customers or suppliers need to change anything?

No. Your business partners can keep sending or receiving PDFs as they do today. PEDIF works with the documents as they arrive.

Does PEDIF use AI?

AI can support the creation of fingerprints. In live operation, known layouts are processed according to the approved fingerprint and rules, so results do not depend on changing model decisions.

Can PEDIF run in our own environment?

In addition to the hosted service, an on-premises option is available for companies with stricter data control requirements. Details are described on the security page.

How do we get started?

You provide sample documents for the document types and layouts in scope. The setup is configured and tested with these samples, you review the results, and processing goes live once the output meets the agreed requirements.