Jeremiah Peino Sakayion group finance lead at FMEC Group, explains why AI could not fix broken financial reporting. He notes that leaders still need humans to translate numbers into business decisions.
This conversation happens in boardrooms and finance functions across Africa and beyond. It usually starts with leaders asking if they can simply use AI for that.
The “that” in question is almost always something that has been broken for years. It could be a reporting cycle that produces numbers nobody fully trusts or a management pack that arrives too late to influence any real decision. It could also be a P&L that tells leadership what happened but never quite explains why. The hope is that AI will walk in and fix the plumbing.
But while AI is a powerful efficiency layer, it is not a cure for structural dysfunction.
The root causes are almost always upstream of technology when financial reporting fails to serve executive decision-making, and they tend to sit in one of three places.
Integrity gaps: Data is often inconsistently defined, incompletely captured, or siloed across systems that were never designed to speak to each other. No algorithm resolves the disagreement when the operations team counts revenue differently from the finance team. It is a governance problem.
Process failures: Reporting workflows are often designed around compliance cycles rather than decision cycles. Month-end closes take three weeks. Reconciliations happen manually in several spreadsheets. Notes exist to satisfy an auditor rather than to generate insight. These are architectural flaws rather than inefficiencies that automation can paper over.
Insight translation gaps: These occur when numbers are technically accurate but operationally meaningless to the people who need to act on them. A CFO who cannot connect a gross margin movement to a specific operational driver is lacking a translator. That translator is the finance function, and this is a human role.
I have seen this pattern repeat itself across every sector and geography I have worked in. The organisations that struggle most with financial reporting are not the ones without sophisticated software. They are the ones where these three underlying tensions have never been properly resolved.
Where AI actually earns its place
AI earns its place as an efficiency layer when the foundation beneath it is sound. It is genuinely transformative in that context, as it can collapse a five-day reporting cycle to a few hours. It can flag anomalies that a human analyst would miss at 11pm on a Friday. It can also surface patterns across years of data in seconds. This gives finance teams more time to think rather than more time to type.
I have helped build robust business processes in various environments. The introduction of AI tools on top of that foundation has been multiplicative. The key phrase is “on top of”. AI amplifies a strong signal, but it also amplifies a weak one.
It cannot build the relationship between the finance function and the commercial team. That relationship makes financial insight actually land in a strategy conversation. It absolutely cannot substitute for a finance leader who understands the business deeply enough to know when a number is telling the truth.
The executive insight problem is a human problem
I observe a specific tension most acutely. Executives do not just need accurate numbers; They need grounded numbers. These figures are anchored to operational reality and delivered with enough context to support a decision.
The gap between financial reporting and executive insight is fundamentally a translation problem. Finance produces outputs in the language of accounting. Leadership makes decisions in the language of markets, customers, operations, and strategy. Bridging those two languages requires someone who is genuinely fluent in both. This person must look at a deteriorating cash conversion cycle and immediately connect it to a customer concentration risk.
This is a talent and positioning problem. It is about where the finance function sits in the organisational structure. It is also about whether finance leaders are given the access and mandate to understand the business deeply enough to serve it well.
What this means for finance leaders
Audit your foundation before you invest in an AI reporting tool. You must ensure your data definitions are consistent across departments. Your reporting processes must be designed for insight rather than compliance. Your finance team needs the operational access and credibility to translate numbers into meaning. The technology investment will disappoint you if the foundation is lacking.
AI is the accelerant you have been waiting for if your foundation is sound. Layer it deliberately. To do this, start with the highest-volume, lowest-judgement tasks. Let it give your team back the cognitive capacity to do what only humans can do, as they need to think.
Do not start with the software if you are building or rebuilding a finance function. Start by defining the decisions the organisation needs to make, then determine what the finance function needs to look like to support them.
That requirement is ancient, and the tools for answering it are evolving rapidly, but the judgement required to execute it well has not changed.
Finance is navigation
I have spent a decade building exactly this kind of bridge. I often sit with operations teams to understand what drives cost behaviour and challenge commercial assumptions before they become embedded in a budget. I also restructure reporting frameworks. This ensures the numbers that land on an executive’s desk carry the weight of genuine insight rather than the hollow precision of uncontextualised figures.
I have worked through these realities by rebuilding broken reporting systems and addressing executive frustrations with unhelpful data. I have found that meaningful change rarely begins with tools. Organisations that will thrive in AI-driven operations are those with clarity on AI’s purpose. They also understand the enduring role of human judgement.
