Look at the map as it stands this spring. Italy has run mandatory clearance for years and extended it to the smallest taxpayers last summer. Poland and France are legislating for phased obligations beginning next year. Spain's framework law has passed with the detailed regulations still pending. The European Commission's proposals on value-added tax in the digital age, published in December, point towards electronic invoicing as the default across the union later this decade and remove the need for member states to seek derogations before mandating it. In Asia, Vietnam completed its rollout last year, India keeps lowering its threshold, and Japan's qualified invoice system starts in October. In this region, the Saudi integration waves are already running and further waves are scheduled through this year and next. For a multinational, that is not a series of local tax updates. It is six or more statutory go-lives inside three years, each with its own deadline and its own definition of what a valid invoice is.
A mandate is not a tax project with an IT component. It is an integration programme with a statutory go-live date you cannot move
This misclassification is the single most expensive mistake being made right now. The work is assigned to tax, because it concerns invoices and the counterparty is a revenue authority. Tax then discovers that the deliverable is a real-time interface to a government endpoint, with certificates, message signing, error handling and a response leg, and that the quality of the output depends on master data owned by three other functions. By then the date is nine months away and cannot be renegotiated.
Four architectural models, and why the difference matters more than the format
Clearance. The authority validates and stamps the invoice before or as it is issued, and an invoice that has not cleared is not legally an invoice. Italy and the Saudi integration phase work this way. Decentralised continuous transaction control. Certified intermediaries or an accredited network carry the document between the parties and report to the authority, as the French design envisages and as Peppol-based systems across Asia and northern Europe already do. Periodic digital reporting. Structured data submitted on a cycle rather than per transaction, often alongside standard audit files. Post-audit with mandated formats. Common for public sector invoicing, where the format is compulsory but no real-time validation occurs. The formats differ, but formats are a rendering problem and every serious vendor solves them. The models differ in something harder: whether a third party must approve your document before your customer owes you money.
Build one invoice object, then adapt
What every regime wants is broadly the same. A uniquely identified document with an unbroken sequence. Complete and currently valid tax identifiers for both parties. Tax treatment expressed at line level rather than derived at the footer. Reliable timestamps. An immutable stored original. Defined treatment of corrections and credit notes. Produce one canonical invoice object that satisfies all of it, and treat each country as an adapter over that object. The organisations that instead let each local finance team commission its own solution end up with five implementations, five support arrangements and no ability to answer a group question about invoice volumes. Be honest about the part that does not generalise, though, because it is where the effort goes: clearance response handling, cancellation and correction rules that differ sharply by country, national identifiers and printed codes, and archiving obligations that may specify where the legal original must physically reside.
The three things that actually break
Master data. Missing, stale or malformed counterparty tax identifiers. In a post-audit world this produced a query months later. Under clearance it produces a rejection in seconds, and the invoice does not exist. Corrections. Most finance teams handle mistakes by cancelling and reissuing. Several regimes do not permit cancellation of a cleared document at all, requiring a credit note with a specific reference structure. Teams discover this in week one of go-live, at volume. The response leg. Every design covers the outbound message. Far fewer cover what happens to the eighty rejected invoices sitting in a queue on a Thursday afternoon, who watches that queue, and what the customer is told. This is an order-to-cash problem wearing a tax costume, and it is where cash gets stuck.
Somebody has to own the queue
Tax owns whether you comply. Technology owns the interface. Neither owns the operational reality of rejected documents, and that gap is filled by nobody until the first bad week. Name an operational owner before go-live, with a service level for clearing rejections, a daily report that someone actually reads, and an escalation path that does not run through a project manager who will be reassigned in June.
Practical Guidance for E-Invoicing Compliance Assessment
- Map every jurisdiction where you issue, including entities finance does not usually think about.
- Classify each regime by model, not by file format.
- Build one canonical invoice object with national adapters over it.
- Validate counterparty tax identifiers now, before a regulator does it for you.
- Decide correction handling per country before go-live, not after.
- Name an owner for the rejection queue, with a service level.
- Check where the legal original must be archived.
- Track your customers' mandates, not only your own.
The Regional Angle
Three issues deserve specific attention from groups headquartered in the Gulf. The first is that the mandate likely to affect you soonest may not be your own. A regional manufacturer, trading house or services firm selling into Italy, Poland or France will be drawn into its customers' regimes: required to transmit through an accredited network, to supply documents in a structured national format, or to be reachable by an access point so that the customer can meet its own obligations. No domestic obligation triggers this and no local tax authority will warn you about it. It arrives as a procurement condition or as an email from a customer's accounts payable team, usually with a short deadline, and the commercial consequence of failing to meet it is that invoices go unpaid. Inventory your customers by their country of establishment rather than by where you invoice from, and start that inventory this quarter, because the French and Polish timetables begin inside eighteen months. The second is an intercompany problem created by the region's own corporate geography. Groups here typically hold a mixture of mainland and free zone entities, several tax registrations, and a substantial volume of intercompany invoicing that has historically been tidied up at period end. Clearance regimes validate registration status at the moment of issue, which converts a tolerated housekeeping habit into a hard failure: an intercompany invoice carrying a lapsed, incorrect or inapplicable registration will be rejected by the authority rather than corrected by an accountant in week three. Use the intercompany population as your first test set. It is internal, so a failure embarrasses nobody outside the group, and it will surface exactly the registration and entity data problems that would otherwise appear on the first day of a customer-facing go-live. The third concerns where the legal original actually lives. A cleared invoice carries a cryptographic stamp and an authority reference, and in several regimes that stamped document, not your rendered copy, is the legally valid instrument. Retention rules across the region and in Europe specify periods, formats and sometimes territory for those records, and they do not all agree. Groups that have consolidated document archiving into a single global platform hosted in Europe or North America should establish, in writing and before the next wave, whether each jurisdiction's original may be held there. This is not a theoretical concern: it determines whether you can produce a valid document in an audit five years from now, and retrofitting an archive split by jurisdiction after the fact is considerably harder than specifying it at the start.
The objection worth taking seriously
The strongest objection is that this is exactly what specialist compliance providers exist for. They track every regime, maintain every format, hold the network accreditations, update when a schema changes and spread that cost across thousands of customers. Building internal capability to do the same is expensive, duplicative, and precisely the kind of programme that turns a manageable subscription into a two-year initiative with a steering committee. That is right, and most organisations should buy connectivity and format maintenance rather than build it. A finance team writing its own signing logic against a national tax endpoint is solving a problem someone else has already solved better. But be clear about what the subscription does not cover. No provider can repair your counterparty master data, decide your correction policy, make your ERP emit line-level tax treatment it never captured, or watch the rejection queue on your behalf. Nor does it transfer liability: when an invoice is not cleared, your customer does not pay, and the authority's question is addressed to you. The right split is explicit — buy the pipes, own the data, own the exceptions, and write down which is which before signing, because that boundary is where these implementations fail.
Common Questions
We only sell domestically. Does any of this apply?
Probably yes, over time. Most domestic regimes are moving in the same direction, and the practical trigger is often a large customer or a public sector contract requiring structured invoicing before any statute does.
Should we wait for the European proposals to be finalised?
No. National mandates are proceeding on their own timetables regardless, and the direction of travel is already clear enough to design against.
How long does implementation take?
For a single entity with clean master data and a capable provider, a few months. For a group with several entities, intercompany flows and inconsistent registration data, plan for a year and start with the data.
What should we expect over the next twelve months?
Expect at least one announced timetable to slip, because they usually do, and plan as though none of them will. Expect the European proposals to shape national designs well before they are adopted, which makes the emerging designs a reasonable guide to what is coming. Expect more Asian jurisdictions to adopt network-based models rather than building bespoke national portals. And expect the first mandate that genuinely affects your business to arrive through a customer rather than through a regulator, which is why the customer inventory matters more this year than the legislative tracker.
E-Invoicing Compliance Assessment — we map your entities and customers against every live and announced regime, design one invoice object with national adapters, and make sure someone owns the rejection queue before go-live.