The governance question nobody asks before approving an AI budget

Somewhere in the last decade, enterprise AI budget conversations developed a rhythm. A vendor presents. Executives nod. Someone asks about the timeline. Someone else asks about the cost. The budget gets approved.

And then, about six months later, the same organisations are asking why the initiative has not moved.

I have seen this in Singapore boardrooms, in Kuala Lumpur, in Sydney and Auckland. The sectors differ. The scale differs. The rhythm is the same.

The answer is almost always the same too. The business problem was never clearly defined. And because it was never clearly defined, nobody could agree on what solving it was actually worth, which resources needed to be committed, or who was accountable when the work stalled.

This is the question that should come first, and almost never does. What specific problem are we solving, and what does it mean for the business if we solve it? That could be revenue growth in an underserved segment. It could be a manual process that absorbs headcount at a scale that has become operationally unsustainable. It could be a customer experience gap that is measurable in churn, or a regulatory obligation that carries real consequences if it is not met. The specifics matter. Broad capability statements are hypotheses.

The second question is how the organisation will actually decide. In a large or matrix organisation across APAC, that is rarely one person. Finance has a view. The business unit has a view. IT has a view. Legal and risk have requirements. In many of the organisations I have worked with across Australia, Singapore, Malaysia and New Zealand, the challenge is not a shortage of opinions. It is the absence of a structure that brings those perspectives together and produces a position that holds. Who convenes that structure? How do they reach a collective decision? Who owns the call when the views diverge?

Microsoft's 2024 AI Business Leaders Adoption Report found that 79% of business leaders see AI as critical to competitiveness. Fewer than half have a defined governance structure for AI investment decisions. That gap is consistent with what I observe on the ground across the region. It is not a technology problem. It is an accountability problem, and the consequences are predictable.

A POC that succeeds technically but never converts to production, because the business case was never tight enough to justify the next commitment. A rollout that encounters problems nobody had the authority to address, because accountability was distributed at the moment the decision was made and impossible to locate when things went wrong.

Getting to genuine AI readiness starts with an internal conversation that most organisations have not had. Three ways to begin it.

Write the problem in one sentence. Who is affected, what is happening, and what does better look like in terms someone can measure. If that clarity exists in a business unit rather than at the organisational level, that is worth knowing early and worth surfacing before the budget moves.

Map who needs to move together. Identify every function with a stake in the outcome and the person in each who can commit their team. Then work out how they produce a collective decision. If there is no clear answer, that is the first thing to build.

Define what governs the work before it starts. What are the risk, regulatory, and operational boundaries this solution must stay within? Who owns those boundaries and who is responsible if they shift?

The organisations that do this well tend to be more precise than fast. They make fewer decisions, but the decisions they make hold.

If this resonates with something you are navigating, share it or leave a comment with what has worked or not worked in your experience. The most useful thinking in this space comes from people who have been in the room.

JR Advisory is opening late April bookings for scoping calls. If your organisation is approaching an AI investment decision and needs structured support before committing, the conversation starts at jradvisory.co. Book an introductory call at: https://calendly.com/jradvisory/initial-discussion

Citation — Microsoft 2024 AI Business Leaders Adoption Report: https://www.microsoft.com/en-us/worklab/work-trend-index.