**Mandatory E-Invoicing Deadline in 2028: 18 Months for SME Accounting Departments**
7 min read
From 1 January 2028, PDF invoices will no longer be permitted in B2B transactions. That may sound far off, but it isn’t. In May 2026, Bitkom confirmed that 41 percent of German companies are already using AI, and one in five SMEs still lacks a digital strategy. Put those two trends together and you’re staring at an 18-month deadline bearing down on mid-market finance teams.
Key Takeaways
- 2028 is the legal cut-off, 2026 is the operational one: ERP and invoice workflows need 12–18 months of lead time. Starting in 2027 simply shifts the risk onto staff already bogged down in day-to-day operations.
- AI without an ERP backbone is just demo theatre: Bitkom reports that 77 percent of AI users see competitive gains-but only where invoices are already structured. PDF clouds stall every AI rollout.
- Grants and mandates align for the last time: Mittelstand-Digital centres offer free advice, and ERP subsidies for 2026 are still available. From 2027 the window narrows-and consultant rates rise.
Related:AI in Accounting: 78 Percent of Invoices Still Entered Manually / S/4HANA Migration 2026
The deadline that’s closer than it looks
The numbers sound dry, but the timeline is brutal. From 1 January 2025, companies must be able to receive e-invoices. From 1 January 2027, mid-sized firms must send them. By 1 January 2028, PDF tolerance vanishes entirely. Many SMEs miss the point: receiving is the easy part; sending forces an ERP overhaul that cannot be improvised. A realistic roadmap runs like this: ERP audit and supplier checks in the second half of 2026, pilot programme with selected suppliers in early 2027, full switch-over by mid-2027. Anyone who hasn’t pencilled this into the investment plan today will face the classic emergency scramble in Q4 2027-complete with overbooked consultants and inflated fees.
AI without an ERP backbone is just demo theatre
Bitkom’s May 2026 figures leave no room for doubt. Forty-one percent of companies now use AI, more than double the share a year ago. Seventy-seven percent of those users report improved competitiveness. Yet dig deeper and a second truth emerges: the gains accrue only where AI is embedded in existing ERP and process landscapes, not in isolated tools. The uncomfortable corollary? A mid-market business still collecting invoices as PDFs in inboxes cannot credibly introduce AI-based booking. The models demand structured data-supplier IDs, tax codes, booking texts in uniform format. If you’re still OCR-ing PDFs in 2026, you’re building the ceiling for your AI ambitions at the same time. The e-invoicing mandate is therefore less a compliance issue and more a prerequisite for AI.
The three investment paths
What I repeatedly observe in consulting sessions in 2026: SMEs face three distinct paths-all valid as long as they’re chosen consciously. Path one: modernise the existing ERP. SAP S/4HANA, Sage 100, Microsoft Dynamics 365 or DATEV Mittelstand Pro can all handle e-invoicing. The effort lies in customising, not licensing. Advantage: no data migration. Disadvantage: legacy processes are preserved, limiting leverage for AI automation. Path two: switch to a cloud ERP. Odoo, Xentral, Weclapp or Microsoft Business Central. Advantage: native e-invoicing, built-in AI modules, faster updates. Disadvantage: migration takes 8 to 14 months; operational responsibility partly shifts to the vendor. Those who can shoulder it gain a full stack reset. Path three: keep the ERP and bolt on an e-invoicing module. Providers like ecosio, Comarch or Basware translate between ERP and invoice formats. Advantage: fastest compliance. Disadvantage: an extra contract, an extra role that can vanish. Best used as a bridge, not a final state.
What the executive team must decide now
Three decisions cannot be delegated because they sit at the SME’s budget and liability frontier. First: who owns the e-invoicing initiative inside the company? Accounting, IT or process owners in sales? In practice only a clear owner with budget and mandate works. Offloading it solely to accounting produces an island solution that stalls AI initiatives. Second: when to launch supplier and customer communication? Suppliers need lead time to adapt their systems. If you work with 50 key suppliers, announce the change in the second half of 2026-or risk your format becoming a brake on supplier relations. Third: which grants will be tapped? The Mittelstand-Digital centres offer free advice. Bavaria, North Rhine-Westphalia and Baden-Württemberg run their own digitalisation programmes with grants up to 50 percent. Failing to draw on these funds means financing the same upgrade later from cash flow.
Frequently Asked Questions
What happens if we still can’t send e-invoices by 1 January 2028?
Input VAT deduction for the recipient becomes more difficult or may be denied-this is the direct financial lever. Reputational costs also loom large, as major B2B customers increasingly reject PDF invoices outright. Fines are possible, but rarely the main issue.
Which formats will be permitted after the mandate?
The decisive standard is the European norm EN 16931. In Germany, XRechnung and ZUGFeRD (from version 2.x) dominate. ZUGFeRD’s strength is combining PDF and structured data in one file-making the transition smoother.
Do we need a new ERP or is a converter enough?
It depends on your ambition. If compliance is the only goal, a front-end converter will suffice. If you want AI-powered booking and automated workflows, you’ll need an ERP that natively processes structured data. Both approaches are valid-just choose consciously.
Which funding programmes will be available in 2026?
Nationwide: Mittelstand-Digital-Zentren offer free consulting. State-specific options include Digital-Bonus Bayern, Mittelstand 4.0 NRW, and Digitalisierungsprämie Plus Baden-Württemberg. Grants range from 5,000 to 50,000 Euro, with formal application phases lasting 8 to 12 weeks.
How long does a realistic migration take?
Existing ERP with module update: 4–6 months including testing. ERP replacement: 12–18 months including data migration and training. Pure converter solution: 6–10 weeks. In all cases, staff availability-not technical setup-is the bottleneck.
Editor’s Reading List
- AI in Accounting: 78 % Dark Posting in SMEs
- S/4HANA or Maintenance Trap: SME Decision 2026
- Process Optimization Without the Marathon Project