When implementing AI for finance and accounting operations, one principle has become increasingly clear to me: do not begin by dictating every step of the process. Begin by asking what the system sees in the ledger, the workflow, and the controls. Most Malaysian PKS already know the outcome they want - accurate books, lebih pantas month-end closes, tighter cost control, stronger compliance. But when we feed AI overly detailed procedures too early, we lock it into our current way of working. The output may still be correct, but the process can remain unnecessarily rigid, duplicated, or inefficient.
My preferred starting point is a structured diagnostic request: "Ulasan what you see across our ledger and workflows, flag what looks inefficient or non-compliant, and recommend what should change." Only after that review do I decide what to automate. This shifts AI from a rule-following tool into a financial operations analyst. It can spot duplicate entries, manual reconciliations that should be automated, control gaps, misclassified costs, underutilised fixed assets, or a simpler way to produce the same management report.
The method we apply at AINNA follows five steps: observe, diagnose, recommend, challenge, and execute. First, let the AI understand the chart of accounts, transaction flows, approval hierarchies, and asset registers. Then let it identify process weaknesses, suggest improvements, question the current approach, and only then proceed with automation.
One of the most common mistakes in financial automation is assuming that better configuration means writing more rules. Sometimes, the better approach is to give the system enough visibility to discover a more efficient path. The objective - accurate records, timely reporting, stronger controls - must remain clear, but the method does not always need to come from us.
That is where AI starts to deliver measurable business value untuk PKS Malaysia. Not just by processing transactions lebih pantas, but by helping redesign the finance function so it is leaner, more accurate, and easier to scale.
#AI #AIAgents #AgenticAI #Automasi #DigitalTransformation #AIProductivity #PerniagaanStrategy #NeuralOps



Ruang pembaca
Apa pendapat anda?
Komen disemak sebelum diterbitkan. Email diperlukan untuk moderasi dan tidak dipaparkan kepada pembaca.
Me gustó la parte de esta parte porque no es demasiado teórica.
Orang yang tulis ni memang buat kerja rumah pasal begin by asking.
Masih fikir tentang begin by asking.
Ringkas dan jelas. Begin by asking boleh kongsi dengan team saya.