Careers of Tomorrow
Within two years, half of the UAE’s government services will run on agentic AI, and Dubai has set its private sector the same horizon. Systems will act for organisations at a speed with no precedent.
Work has changed location. Accountability has stayed with people. This page is about what that asks of the professionals who run modern businesses, and how the answer is already taking shape.
When systems act, three questions remain with people
For most of the modern enterprise’s history, a decision had a buffer around it. A transaction happened, it was recorded, someone reviewed it, someone else reconciled it, and a report reached leadership weeks later. The buffer was where judgement lived. It was where errors were found, where interpretations were debated, and where a professional earned the right to sign.
That buffer is closing from two directions at once. Regulators receive structured transaction data at the moment of the transaction. Enterprises are building finance functions that see the business as it happens. And the UAE’s agentic-AI programmes for government and for Dubai’s private sector give both directions a two-year clock. In each case the machine executes, the record is created at source, and the window for later correction shrinks towards zero.
Speed changes how business gets done. It also changes how professionals think, because three questions stay with people however fast the systems move.
Who still understands the outcome
when execution is spread across systems, models and teams?
Who owns the decision
when the system produced it from rules someone approved months ago?
Who can explain and defend the result
when it is questioned later, by a board, a regulator or a customer?
Those three questions define the careers of tomorrow. The rest of this page follows the evidence: what enterprise finance is building, what tax administrations have already built, and what both are telling us about the professional the future needs.
Seeing the business as it happens
The clearest public account of an AI-native finance function comes from OpenAI’s Chief Financial Officer, Sarah Friar. Her stated goals are a zero-day close, described as “a real-time, reconciled, and traceable view of the company’s financial position”, and continuous forecasting, “a continuously refreshed view, faster scenario analysis, and clear finance ownership”. Spending plans, purchase orders, actuals and accruals are reconciled continuously. AI drafts the variance explanation. The finance professional validates it and owns it.
“AI accelerates the work. People own the result.”
Sarah Friar, CFO, OpenAI
Three things in that account matter beyond OpenAI. First, visibility attaches upstream, at the commitment and the purchase order, well before the invoice. Second, the rules that drive the numbers, the forecasting baselines and allocation logic, are governed: changes need finance authorisation. Third, success is measured by whether the result was usable and whether it enabled a faster or better decision, in place of licence counts and adoption metrics.
Each of those three has a precedent. Tax administrations reached the same conclusions years earlier, at national scale, against legacy ERPs and multi-entity groups.
Source: Sarah Friar, “What building an AI-native finance function taught me”, OpenAI, 2026. Quoted as published evidence of the direction of enterprise finance.
The completed experiment
Tax administrations built the first zero-day close, and they built it for the whole economy. Under Tax Administration 3.0, the OECD’s description of tax embedded in the systems businesses already use, an invoice reaches the authority as structured data at or near the moment it is issued. Brazil, Italy, India and now the UAE under PINT-AE run continuous transaction controls on this model, and the PEPPOL-based regimes spreading across the GCC and Europe extend it further each year.
That makes real-time tax the most complete evidence available on what happens when an observer sees a business at transaction speed. Five lessons stand out, and each one transfers directly to the finance function now building the same capability for itself.
- 01
Semantics come before speed.
The authority’s view became trustworthy once a shared document standard fixed what each field means. A finance function that accelerates before it fixes a canonical transaction semantic gets faster disagreement.
- 02
Validation at source beats repair downstream.
Clearance regimes validate before the invoice enters the ecosystem. A zero-day close depends on the same discipline applied at the purchase order.
- 03
Speed exposes interpretation.
Mismatches that used to surface at quarter-end now surface within hours. The interpretation behind a number becomes visible sooner, and it needs an owner.
- 04
Speed without evidence creates a defensibility gap.
A cleared invoice with a weak evidence trail is a liability the moment it is questioned. A forecast a CFO can act on and trace is the finance equivalent.
- 05
When machines do the work, accountability has to be written down.
It used to fall out of who performed the task. It now has to be designed.
Tax Administration 3.0 explained →The Global E-Invoicing Tracker →
A vocabulary for the real-time enterprise
These frameworks were developed in practice, inside enterprise tax and ERP transformation, and they describe the general problem of governing decisions that systems now execute. Each links to its full definition.
The closing interval between a transaction and the authority’s structured visibility of it.
How fast must internal understanding move once external understanding is instant?
The continuous view of a business that an observer assembles from its structured data.
What does leadership’s own mirror of the business need to be built from?
A technically green status, transmitted and validated, that masks an incorrect or incomplete position underneath.
How does an organisation tell a reconciled number from a correct one?
Automated execution that leaves humans accountable for outcomes they can no longer explain.
Who understands the result when responsibility spans several teams and a model?
Approved interpretation translated into executable, version-controlled system logic.
Which rules are authorised to run the business, and who governs changes to them?
The governed proof layer that makes an executed position provable.
Can the enterprise reconstruct the reasoning behind a number it acted on?
Controls embedded at transaction origin, in ERP logic, master data and workflow.
Where does control sit when the window for downstream review has closed?
Read together they describe three clocks: the transaction clock, the authority clock and the management clock. Real-time tax closed the distance between the first two. The AI-enabled finance function is closing the distance between the first and the third. The governing principle for the enterprise follows:
No external party should understand your transaction before you do.
The professional the real-time enterprise needs
The answer to the three questions is a person, and the person has a name. A Fusion Professional™ is accountable for outcomes generated at transaction speed, across systems configured by others, in a data ecosystem an external observer sees clearly. They connect the dots across law, systems, data and governance. Specialists go deep within each domain, and the Fusion Professional™ is the one who can read across them and own the result.
Operator
Executes the process.
Owner
Accountable for its output and for the interpretation behind it.
Builder
Designs the logic, the controls and the evidence that let a system execute correctly.
The career path runs Operator → Owner → Builder™. Value has moved along that path in tax, and the same movement is now visible in finance, in audit and in every function where a system acts on approved rules. This is a reallocation of professional value from downstream rework to upstream design and judgement. The tax technical knowledge remains essential. The role now asks for more alongside it.
Fusion Professionals™ was introduced in Extinction of Tax As We Know It (2025).
Understand it. Become it. Build it.
The conversation is wider than tax
Regulators, CFOs, boards and educators are all asking the same three questions about a world where systems act. Nitin speaks and writes on how businesses keep judgement, accountability and trust when execution moves into systems, and on the capability that professionals and organisations need to get there.
The Real-Time Tax Briefing, monthly for finance and tax leaders, is in the footer below.