In most steering committee kickoffs for a UAE e-invoicing programme, someone will ask whether this is a tax project, an IT project, or a finance project — and the meeting will then spend the next forty minutes on the wrong question.
The question surfaces because the budget, the deadline, and the operational impact each sit with a different function. Tax sees a compliance obligation with a legal deadline and assumes ownership by default. IT reads the same programme as a system integration project with a defined scope — connect the ERP to an accredited service provider, map the fields, pass the schematron validation, done. Finance treats it as a change to the invoice-to-cash cycle, an operational rollout for the shared services team to absorb alongside everything else already on the calendar. With UAE ASP appointments due by October 2026 and go-live mandatory from January 2027, organisations spending their first year answering the ownership question rather than designing around it lose time they will not get back.
The better answer is that e-invoicing sits closest to tax — but the word "tax" in that sentence now carries a different job description than it did a decade ago. Under periodic reporting, tax could reasonably describe itself as the function that interprets law, prepares returns, and defends positions on audit, work that happened downstream of the transaction, on a data set someone else had already created, structured, and posted. Tax did not need to own the underlying data or trace every accounting entry personally, because there was time between the transaction and the authority's view of it in which someone could catch and correct a problem. That division of labour was a sensible response to how much time the old compliance cycle gave a business.
The Interval That Disappeared
Real-time e-invoicing removes that interval. When the authority sees the transaction at or near the moment it happens, there is less room left in which to discover, correct, or explain a data problem before the authority has already formed its own view of it — what the Authority Mirror View™ describes directly: the authority's copy of the enterprise's tax position is built from the same data the enterprise's own systems produced, at the moment those systems produced it. The distance between what the authority knows and what tax knows stops being a filing-season inconvenience once that interval closes, and becomes one of the larger sources of exposure a real-time tax function carries.
India's experience running the world's largest continuous transaction control system gives this a measurable shape. The ClearTax State of Tax Assurance Report 2026 found that sixty-three per cent of CFOs rate their organisation's compliance as strong, while only thirty-three per cent actually have real-time visibility into the data producing that compliance — a thirty-point gap between confidence and the visibility that would justify it. India's tax administration runs AI across its filings today, through infrastructure that by most accounts performs as designed. The confidence gap points instead to tax functions still organised and staffed for a compliance cycle the underlying systems have already outgrown.
What the Right Answer Requires
Saying tax owns e-invoicing only works if tax now means data architecture, system configuration, exception governance, and control design — competencies becoming part of the tax discipline in something like the way treaty interpretation became one a generation ago, sitting alongside the older skill set rather than replacing it. Fusion Professionals™, the role introduced in Extinction of Tax As We Know It, is what makes this operational: a specialist who treats tax interpretation and system behaviour as one connected discipline, because in a real-time environment they already function as one.
The steering committee that spends its first year resolving who owns the project has confused a staffing question with a capability question. The capability the programme needs to build spans tax interpretation, system configuration, data governance, and process ownership, and no single function holds all four alone. Naming tax as the closest owner only works once tax is staffed and trained to hold that span. Building that capability, not settling the ownership debate, is the work that determines whether the January 2027 go-live produces a defensible tax function or a technically connected one.