Building the system is a CFO skill now, not an IT one
Who should own the ERP, finance or IT? The case for the CFO owning finance-systems architecture, the build-versus-buy call, and the controls, from someone who designed and shipped a multi-company ERP in Saudi Arabia.
For thirty years the division was clean. Finance owned the numbers. IT owned the systems that produced them. A CFO who could read a balance sheet did not need to know how the ERP behind it was built, and nobody expected otherwise.
That division is breaking, and the finance leaders who see it are pulling away from the ones who do not.
Why the split stopped working
Every number a CFO signs now comes out of a system. The close runs on it, the VAT filing depends on it, the ZATCA integration lives in it, and the board pack is only as good as the data model underneath. When the system is wrong or slow or cannot answer the question, the CFO does not get to blame IT. The board asked finance, and finance owns the answer.
A CFO who cannot shape those systems is limited to reading whatever the software hands them. If the ERP cannot produce margin by project, they do not have margin by project. If the chart of accounts was designed by a vendor who left three years ago, they inherit its blind spots. Depending on someone else to build the machine that produces your numbers means depending on someone else for the quality of your own work.
What "building the system" means
This is not a call for CFOs to learn to code. It is narrower and more useful than that.
It means owning the architecture: the chart of accounts, the entity structure, the data model that decides which questions the business can answer at all. It means making the build-versus-buy call with judgment instead of taking whatever a vendor recommends. It means specifying what the system must do, governing how it gets built, and knowing enough to tell when an engineer or a vendor is solving the wrong problem. The engineers build. The CFO decides what gets built, and why.
Done well, this is the difference between a finance function that waits two weeks for a report and one that answers the question in the meeting.
The proof is what you can produce on demand
I take this position because I have lived it. I designed and shipped a multi-company operating system for a group in Riyadh: general ledger, AR and AP, VAT, HR, payroll and GOSI, with governance and audit logging built into every module, in English and Arabic. Engineers wrote the code. I owned the architecture, the build-versus-buy calls, and the controls. The result is a finance function that produces its numbers instead of chasing them.
That is the standard the next generation of finance leaders will be measured against. Not whether the statements are correct, but whether the machine that produces them was built to answer the questions the business asks.
Where this goes
Automation and AI widen the gap, they do not close it. OCR reads the invoices, a model categorises the ledger, anomalies surface before the close. A CFO who knows how to deploy that safely, with the controls an auditor accepts, compounds the whole team's output. A CFO who treats it as an IT project watches from outside the room.
The finance leader who can build the machine that produces the numbers is worth more than the one who can only read them. In a market rebuilding itself around Vision 2030, that gap will decide who runs finance for the companies that matter.
If you are building the finance function that has to scale, tell me where you are.
A point of view from experience, not advice on any specific system or vendor.