What AI Actually Changes About Management
- Raffles Jakarta
- Aug 3
- 8 min read
There is a new silence in many meetings. A manager asks a question that would once have taken an analyst two days to answer, and a machine returns a fluent, confident, well-formatted, many-response answer in about nine seconds. The chart is drawn. The summary is written. The forecast is produced.
For anyone whose authority rested on being the person in the room who had the answer, this is a quietly destabilizing moment, because the answer has just become abundant, and abundant things lose their price.
It is tempting to read the situation as a story about managers being replaced. It is closer to the opposite. What is being commoditized is the production of answers. What is becoming scarce, and therefore valuable, is the judgment to know which questions are worth asking, whether the answer in front of you is any good, and what to actually do about it.
That shift is the most important thing happening to the practice of management right now, and most of the language around it obscures rather than clarifies. It is worth looking at what the evidence actually says.
Everyone has the tool. Almost no one has the advantage.
The headline numbers on adoption are, by now, almost boring. In McKinsey's most recent global survey on the subject, 88 percent of organizations reported using AI in at least one business function, up from 78 percent a year earlier (McKinsey & Company, 2025). On the surface, the situation looks like a settled transformation.
Read one layer down and the picture inverts. Of those same organizations, only about a third reported having scaled AI across the business rather than running it in isolated pilots, and only a small minority, in the region of five percent, could point to a meaningful impact on earnings from all of this activity (McKinsey & Company, 2025).
So the tool is everywhere, but the advantage is almost nowhere. That gap is the key point, and it is not a technology gap. The organizations capturing value were not the ones with better models; everyone has access to broadly the same models.
They were the ones that had done managerial work: redesigning how the work actually flows, putting senior leaders in charge of governing the technology, and rebuilding processes around the tool instead of bolting it onto processes designed for a world without it.
The differentiator, in other words, is not the machine. It is the quality of the management wrapped around the machine.
The value was never in the answer. It was in the question.
Three economists articulated the clearest way to understand why in a book that has become the standard reference on the subject. Their argument is that artificial intelligence, stripped of the mysticism, represents a dramatic fall in the cost of one specific thing: prediction.
When the cost of any input falls far enough, two things happen. Demand for that input explodes, which is the adoption story above. And, less obviously, the value of its complements rises. The great complement to cheap prediction is judgment, the human work of deciding what you are trying to achieve successfully, what a successful outcome looks like, and what to do once the prediction arrives (Agrawal, Gans & Goldfarb, 2018).
A machine can tell you, with increasing accuracy, which customers are likely to leave. It cannot tell you whether keeping them is worth the cost, which ones you want to keep, or what trying would do to the business. It can forecast demand. It cannot decide how much risk the organization should carry into that forecast. As prediction becomes cheap, these questions of judgment do not disappear.
They become the entire job. This is why the World Economic Forum, surveying more than a thousand of the world's largest employers, finds AI and data skills the fastest-rising capability employers expect to need and keeps analytical thinking at the top of the skills it calls essential today (World Economic Forum, 2025).
The two findings are not in tension. They describe a manager who sits precisely at the junction of the two: fluent enough in the machine to use it and clear enough in their reasoning to know what to ask of it and what to ignore.
This is an old finding wearing new clothes
None of this began with generative AI, and remembering that helps separate the durable point from the hype. More than a decade ago, well before anyone was talking to a chatbot, researchers at MIT and Wharton examined 179 large public companies and asked a simple question: did the firms that made decisions on the basis of data and analysis actually perform better than those that ran on instinct and hierarchy? They did. Firms that adopted a genuinely evidence-led approach to decisions showed output and productivity roughly five to six percent higher than their investments and technology alone would predict, with the advantage also visible in asset utilization, return on equity, and market value (Brynjolfsson, Hitt & Kim, 2011).
Five to six percent sounds modest until you compound it across a decade against competitors who are guessing. The point that survives into the present is this: the capacity to manage by evidence was a measurable competitive advantage long before the tools were good, and the tools getting good does not remove the requirement.
It raises it. Thomas Davenport and Jeanne Harris made the same case for analytics as a leadership discipline rather than a technical function, arguing that competing on analytics is a strategic choice made at the top of an organization, not a capability that can be delegated downward to a technical team and forgotten (Davenport & Harris, 2007).
What has changed since they wrote is only that the tools have become extraordinary. What has not changed is that a tool cannot make the strategic choice for you.
What "AI literacy" means for a manager, and what it does not
The phrase gets used loosely, so it is worth being precise about what a manager actually needs, because it is not what the word "literacy" usually implies. It does not mean learning to code. It does not mean building models or understanding the mathematics inside them.
A manager who tries to become a junior data scientist has misunderstood the assignment. The literacy that matters is the literacy of interrogation, which separates into four capabilities that we can name plainly.
The skill is reading an output critically rather than producing one. The skill is not generating the analysis; the machine does that. The skill is knowing what a confident answer hides, which assumptions are buried in a metric, and how to distinguish a genuinely informative result from one that is merely fluent. A model will state a wrong answer with exactly the same composure as a right one.
Framing the problem before the tool touches it is essential. The quality of any answer is capped by the quality of the question. Deciding what to actually optimize for, what to measure, and what a satisfactory outcome would even look like is upstream managerial work that no tool performs for you, and getting it wrong means receiving a precise answer to the wrong question.
Governing the risk. Bias in the data, plausible fabrication, and the slow erosion of judgment that comes from over-relying on a confident machine are now management responsibilities. The organizations capturing value are the ones whose leaders took ownership of the process rather than leaving it to chance.
Deciding under uncertainty. The machine narrows the uncertainty; it does not remove it, and it never carries the accountability. Someone still has to commit the organization to a course of action and answer for it. That someone is a manager, and the decision is the part of the job that does not automate.
Notice that none of these are technical skills in the narrow sense. They are the skills of reasoning, framing, and deciding, applied to a new and powerful instrument. This is why the manager who thinks the answer to AI is to hire more engineers has missed the point, and the one who treats it as a challenge to their judgment has understood it.
The part organisations keep getting wrong
There is a supply-side failure here that is worth naming because it explains why so many capable managers feel behind. The same research that documents near-universal adoption also finds that employees are, quietly, using these tools far more than their leaders realize, while the value only materializes where senior people take active ownership of the technology and redesign work around it (McKinsey & Company, 2025). Put those two findings together, and the diagnosis is uncomfortable.
The usage is happening from the bottom up. The value depends on capability at the top. And most organizations are investing in the tools while underinvesting in the one thing that turns those tools into results: the judgment of the people meant to lead with them.
For an individual manager, that is an awkward observation and a clarifying one. If the organization buys the technology but not the capability to lead it, you must arrange that capability for yourself before the gap between what the machine produces and what you can do with it becomes the ceiling on your career.
Where a manager learns to think with the machine, not against it
This scenario is where structured study earns its place, and we should be clear about what it does and does not offer. No course turns a manager into a data scientist, and none should try.
What a serious postgraduate business education does is build the reasoning that sits above the tool: the ability to read financial and operational numbers with suspicion, to understand the economics of a decision well enough to know when a confident forecast is confidently wrong, and to frame and defend a strategy in a room full of people who do not share your assumptions. That is precisely the architecture of the Raffles Jakarta MBA. Information systems management and strategy address technology directly as a matter of strategy, not plumbing.
Managerial Economics and Accounting for Decision Making builds the quantitative reasoning that enables a manager to interrogate an output rather than simply accept it. Strategic Management develops the judgment that the machine can inform but never supply.
It is a one-year Master of Business Administration program taught entirely in English on Jalan M.H. Thamrin in Central Jakarta, with four intakes a year, in January, April, July, and October, and a choice of studying on campus or in a hybrid format that lets working professionals continue their careers while they study.
Thinking bigger, in an age of capable machines, is not about competing with them at what they do best. It is about becoming better at the thing they cannot do at all.
The questions do not automate
The machines will keep improving at producing answers, and that trend is not going to reverse. The mistake is to assume that the trend diminishes the role of the useful manager. It does the opposite. When answers are cheap, the scarce and valuable work moves to the edges that the machine cannot reach: deciding which questions deserve asking, judging whether an answer is useful or good, carrying the risk of acting on it, and standing behind the decision when it is made. Those have always been the real content of management. The arrival of intelligent tools has simply stripped away everything else and left the core of the job exposed, more clearly than at any point before.
The managers who understand that are not worried about the machines. They are learning to think bigger than them.
THINK BIGGER, READ MORE
Marketing Manager
References
Agrawal, A., Gans, J., & Goldfarb, A. (2018). Prediction machines: The simple economics of artificial intelligence. Harvard Business Review Press.
Brynjolfsson, E., Hitt, L. M., & Kim, H. H. (2011). Strength in numbers: How does data-driven decision-making affect firm performance? https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1819486
Davenport, T. H., & Harris, J. G. (2007). Competing on analytics: The new science of winning. Harvard Business School Press.
McKinsey & Company. (2025). The state of AI: How organizations are rewiring to capture value. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
Raffles Jakarta. (2026a). Master of Business Administration. https://www.raffles-indonesia.com/mba
Raffles Jakarta. (2026b). Think bigger: The Raffles Jakarta MBA 2026. https://www.raffles-indonesia.com/think-bigger-mba-jakarta
World Economic Forum. (2025). The Future of Jobs Report 2025. https://www.weforum.org/publications/the-future-of-jobs-report-2025/
