PDF vs structured invoice

The distinction that matters in Belgian B2B

For decades, the invoice in business imagination has been a page: a layout with a logo, line items, and totals. PDF preserved that mental model in digital form. Email made distribution cheap. Together, PDF-by-email became the default channel for Belgian B2B invoicing.

From 1 January 2026, Belgium requires many VAT-liable businesses to exchange structured electronic invoices for domestic B2B transactions — not merely to send a readable document. The mandate concerns the format and delivery channel, not whether humans can still read a summary. A PDF that looks correct on screen does not, by itself, satisfy what the obligation targets.

This article explains why that gap exists, what "structured" means in practice, and how PDF fits as a companion rather than a substitute. For the legal scope of the mandate, see the Belgian structured e-invoicing obligation (2026).

Why PDF-by-email falls short of the mandate

A PDF is primarily a presentation format. It encodes visual layout: fonts, coordinates, perhaps embedded images. Software can extract text, but extraction is probabilistic. Two PDFs from different suppliers rarely share the same field names, table shapes, or VAT breakdown conventions. Each inbound document becomes a small parsing problem.

Belgian B2B structured e-invoicing addresses a different requirement: machine-readable interchange between accounting systems. When the mandate applies, the invoice must circulate as structured data — typically EN 16931 semantics in UBL XML, delivered over the Peppol network — so the buyer's system can ingest seller, buyer, lines, VAT categories, and totals without re-keying or OCR.

Sending only a PDF when structured exchange is required creates process risk: the counterparty may reject the invoice, delay payment, or treat it as non-compliant. Email delivery also lacks the network-level delivery semantics that Peppol provides — a separate topic from whether the PDF opened correctly in someone's inbox.

What "structured" means

"Structured" does not mean "organised on a page." It means the invoice exists as defined data elements with agreed meaning across systems.

At the European level, EN 16931 specifies which business information an electronic invoice must carry: parties, identifiers, invoice number, dates, monetary totals, VAT breakdown, payment instructions, and line-level detail. UBL (Universal Business Language) is a common XML syntax that expresses those elements as a strict tree: AccountingSupplierParty, InvoiceLine, TaxTotal, and so on.

On Peppol, BIS Billing 3.0 profiles EN 16931 for network exchange. A Belgian structured e-invoice in this context is not a free-form file; it is a document that validators can check against mandatory fields and cardinality rules before delivery.

Structured invoice

An electronic invoice whose business content is encoded in a standardised data model (such as EN 16931), independent of visual layout, so software can process it without interpreting a page image.

For the technical stack in more detail, see EN 16931, UBL, and the structured invoice.

The human PDF as companion copy

The mandate does not erase the need for human-readable invoices. Many workflows generate both: the structured XML (or equivalent) as the authoritative e-invoice, and a PDF rendering for people who approve payments, attach files to cases, or archive for quick reference.

The relationship is asymmetric:

• Structured file — source of truth for compliance, automation, and audit of electronic originals
• PDF — convenience copy; useful, familiar, but not a replacement when structured exchange is required

Treating the PDF as sufficient because "everyone can open it" confuses legibility with interoperability. A supplier might email a PDF and believe the job is done; a buyer subject to the mandate still needs structured data to book the transaction without manual intervention.

Accounting automation and the limits of OCR

Finance teams adopted OCR and AI extraction to reduce manual entry from PDFs. These tools can work well for homogeneous document streams, but they remain inference layers on top of presentation. Confidence scores, exception queues, and human correction are normal. Costs scale with volume and supplier diversity.

Structured invoices invert the model. Fields arrive already labelled: VAT category codes, unit prices, quantities, and party identifiers map directly into chart-of-accounts rules and three-way matching. Exceptions drop when the document fails validation upstream rather than after posting.

For businesses receiving high inbound volumes — distributors, retailers, shared service centres — the difference is operational, not cosmetic. The Belgian push toward structured B2B exchange aligns tax compliance with how modern ERP and AP automation are designed to work.

Delivery proof and audit posture

Email proves that a message was sent, not always that the invoice as a business document was received, parsed, and accepted into the buyer's process. Attachments bounce, spam filters intercept, and mailbox rules misfile. Disputes about "we never got your invoice" are harder to resolve from SMTP logs alone.

Peppol delivery introduces network status: submission through an Access Point, routing to the recipient's Access Point, and technical acknowledgements within the exchange framework. That does not replace commercial acceptance of line items or prices, but it strengthens the evidence chain for electronic delivery — relevant for internal control, supplier relations, and demonstrating that structured exchange occurred.

Retention practice should favour the structured original (validated XML or equivalent), not only the PDF snapshot. Auditors and systems replay structured data; PDFs are secondary exhibits.

Practical posture for Belgian businesses

When the Belgian obligation applies to a transaction:

1. Send structured invoices over Peppol (or the channel defined for the mandate), not PDF-only email
2. Receive structured inbound invoices into accounting or AP workflows
3. Offer PDF as a companion where helpful, generated from the same source data
4. Archive the structured original alongside human-readable renderings

PDF remains a durable format for reading. It is not, when the mandate applies, a durable substitute for structured B2B e-invoicing.