In 2029 a PDF stops being an invoice, and the decisions that matter happen this year

From April 2029, an invoice attached to an email as a PDF will stop counting as an invoice for most business-to-business billing in the UK. The government confirmed the mandate in November 2025, and in June this year it named the plumbing: invoices will travel as structured, machine-readable documents over a network called Peppol, from your software to your customer's software, with no human opening an attachment in between.

Three years out, the vendor articles are already selling urgency. The honest position is closer to the opposite. There is nothing you should buy this month, and anyone pushing an "e-invoicing solution" at a ten-person firm in 2026 is early by design, not by accident. What 2026 actually asks of you is smaller and cheaper: stop making the decisions that will age badly.

What was actually decided, and what was not

Two announcements matter. In November 2025 the government confirmed that electronic invoicing becomes mandatory for business-to-business invoices from April 2029, with a phased start that reaches larger businesses first. On 23 June 2026 it confirmed Peppol as the network the whole thing runs on, which is the same framework much of Europe already uses, and said the initial mandate is invoice exchange only: the invoice goes to your customer, not to a new government inbox.

The detailed roadmap, including the phasing and the fine print of who is in scope from which date, is due at the Budget on 28 October. Which firms fall inside the mandate is tied up with how a business is registered for tax, and that boundary is a conversation for your accountant, not a thing to take from a blog, this one included.

April 2029: Structured e-invoicing becomes mandatory, phased
28 Oct 2026: Implementation roadmap due at the Budget

What is already certain is the direction: the invoice stops being a picture of a document that a person reads, and becomes data that a system reads. That single change is what everything else follows from.

Which end of the market you are on

The useful question in 2026 is not "what do I buy". It is "which of two groups am I in", because they face completely different futures.

If your invoices come out of a mainstream ledger, Xero, QuickBooks, Sage and their peers, you will most likely be carried across by your vendor. Peppol support becomes a feature of the software you already run, the way bank feeds did. Those vendors sell in a dozen countries that already mandate e-invoicing, so the capability substantially exists; what remains is switching it on for the UK flavour and charging you accordingly. Watch the pricing page, not the technology.

If your invoices come out of Word, Excel or a template someone built years ago, you are in the other group, and the mandate is a real project with your name on it. Not this year. But the gap between the two groups is exactly the spreadsheet-to-system gap we have written about before, and the mandate has quietly put a deadline on one specific corner of it: by April 2029, at whatever phase catches you, the invoicing part of the spreadsheet has to have become a system.

The decisions that age badly

Here is the part the urgency-sellers get right for the wrong reasons: some choices made in 2026 will look expensive in 2028.

Building new process around unstructured invoices is the big one. Every workflow you construct this year that depends on a PDF landing in an inbox, being read by a person, and being retyped into something else, is a workflow with a demolition date on it. If you are designing an invoicing or invoice-handling process right now for any other reason, design it around structured data leaving one system and landing in another, because that is the shape everything is being pushed towards anyway. Firms already doing this get the mandate for free.

Commissioning bespoke invoicing is the second. A custom-built invoicing module specified in 2026 that does not have Peppol on its roadmap is a liability being written to order. We build bespoke systems for a living, so this is not an argument against bespoke; it is an argument that "how does this reach Peppol in 2029" now belongs in the specification of anything that raises or receives invoices, ours included.

The third is quieter: signing a multi-year software contract without asking the question.

01At your next accounting-software renewal, ask the vendor what their Peppol plan is for the UK mandate and whether it will cost extra
02If any system you commission or integrate raises or receives invoices, put the 2029 requirement in the specification now
03If your invoicing lives in Word or a spreadsheet, put the move onto a proper ledger on next year's plan rather than 2028's

Why waiting is also a decision

The temptation on reading all this is to file it under 2028's problems, and for buying software that is right. But the firms that will find 2029 expensive are not the ones that bought nothing in 2026; they are the ones that spent the next two years accumulating process the mandate will tear up. Doing nothing with your systems is a choice like any other, and it compounds.

Our sister practice, Opus Accountancy, sees the invoice formats small firms actually send, which is how we know the Word-and-spreadsheet group is bigger than the software industry likes to admit. And the platform we built for them runs a live Xero integration, so we have working knowledge of what it takes for an invoice to move between systems as data rather than paper. If you want to know which group you are in, or what the Peppol question means for a system you are about to commission, ask us; in 2026 that conversation is short and cheap, which is rather the point.

We work with businesses across Knutsford, Alderley Edge, Wilmslow, Altrincham, Stockport and Warrington, and remotely for clients anywhere in the UK.