We now introduce our discussion on strategic efficiency. To do so, we first need to understand what data can do and what data cannot do. The best way to tackle this problem is by discussing the amazing developments that we observe today in terms of availability of data and elaboration capability. Erik Brynjolfsson and Andrew McAfee have studied the impact and implications of artificial intelligence for business activities. They highlight that AI has improved dramatically in two areas: voice and visual recognition, and in addressing problems that can be analyzed by using very large and fine datasets. For example, companies can predict where to allocate their sales forces as a function of where demand peaks in the future or how to optimize production based on millions of data points about production activities and performance. As a matter of fact, the typical applications are in areas where large and fine data are available, such as marketing, operations, human resources. Machines are moving to higher-level decisions as well. However, it's not unfair to say that apart from some frontier applications, in quite a few higher-end areas, the human manager is still the central figure. In particular, managers know how to frame the problem, how to understand causal relations, and how they can make generalizations. These are crucial elements of any decision-making process, at least of less standard ones. There are two ways of thinking of what data alone cannot do. The first one is that data can only explain the world as it is, but they do not imagine how the world could be, as Roger Martin and Tony Golsby-Smith had put it in a 2017 article on the Harvard Business Review. They make the example of Lego's CEO, Knudstorp, who was puzzled by the fact that 85 percent of the kids playing with Lego were boys. He imagined how the world could be by changing some key features of the Lego bricks to produce Lego friends, that have become far more popular with girls. Similarly, László Bíró figured out that a rotating ball on the tip of a pen could pick the ink and put it on the paper, overcoming the problem that sticky ink does not flow down. But even László Bíró did not imagine the whole thing. He thought of his pen as a luxury good. A little later, the Frenchman Marcel Bich, realized that if the pen was made cheaply in plastic, he could profit by selling many pens all over the world. The second point is that data cannot tell us what to do with them. Sure, if we have an excel file listing, for example, prices and quantities or other measures, we can compute means or show some graphs. But without some good ideas about what to do with them, the data by themselves are not very informative. Similarly, huge datasets from LinkedIn, or millions of observations, for example on electricity consumption at anytime of the day of a region or country, are clueless unless we know what to predict and why, or how to use them more generally. Management scholars or courses like this one cannot teach you about the first problem. We are not experts in your domains and thus we cannot teach an entrepreneur or a manager how to get a good idea. You have to develop your own good ideas. You have to imagine how your world could be. However, we can teach you about the second problem. We can teach you how to assess an idea that you have and how to use data and other tools for this purpose.