Toshiyuki Ueyama, Director of Chiba University of Commerce Research Center for Economics

Recently, the term "IoT" has become more common in newspapers and magazines. Of course, it stands for Internet of Things, and simply put, it means connecting various things to the Internet to exchange information, but until recently the word "ubiquitous" was often used, and terms like "ubiquitous society" were also in use. The two terms IoT and ubiquitous have different purposes and nuances, but if you look past the details, you can think of them as being the same thing for now. Incidentally, starting to use a new word can have a positive impact on business in related industries.

Now, IoT is being discussed at the Ministry of Economy, Trade and Industry, and reports have been released, so you can find out what discussions are currently taking place on the web. The most well-known example of IoT is undoubtedly KOMTRAX, a machine operation management system built by Komatsu Ltd., which remotely collects and manages data from the company's construction machinery, and even provides guidance on how to improve energy-saving operation.

As a similar analogy, around 1990, shortly before the rapid spread of the Internet, Coca-Cola Japan's bottlers began experimenting with collecting inventory status data from vending machines via radio and telephone lines in order to improve the efficiency of product replenishment by route sales by implementing a "one-way operation method" for the replenishing of products in vending machines, and later introduced the system. Therefore, while there have been cases in the past where data was collected remotely to improve the efficiency of operations, we can now say that we have the means to achieve this relatively easily and inexpensively.

Furthermore, efforts are currently being made to use AI (artificial intelligence) to generate information and knowledge from so-called big data after data is collected via IoT.

Around the 1990s, AI research was active, but at the time, research was centered on rule-based artificial intelligence, where humans would have computers memorize facts (data) and rules and then make inferences. The author also once worked on a prototype expert system that could perform consulting by incorporating knowledge gained from his experience in management consulting. However, today's AI is completely different from that time in that computers are capable of self-learning. It might be fair to say that we are at the beginning of an era in which computers self-multiply knowledge. There is also the aspect that IoT, big data, and machine learning will progress together.

Here, I would like to think about management separately from that discussion. Of course, the development of IoT, big data, and machine learning as a set is important, but I would like to go back to the basics and think about management and business.

Generally speaking, new business opportunities arise when a new system is introduced or changed. The recent introduction of the My Number system is also a major business opportunity. However, when changes occur gradually, rather than as an explicit change where a new system is implemented at a set time, it can be difficult to notice them, or, as is often the case with Japanese companies, even if they do notice them, they may overlook them. Naturally, in such cases, management is often at a disadvantage. As for IoT, as mentioned above, it is already widely known, so there is a sense that starting now would mean missing the trend, but there are still companies that need to start working on it now.

What happens when you think about IoT in relation to the development of your own products and services? First, IoT can be broadly categorized into cases where the emphasis is on data collection using IoT, cases where the emphasis is on data distribution, and cases where both are done. Cases where the emphasis is on data collection include systems that monitor weather conditions near railway tracks where it is difficult for people to patrol, and systems that collect data on manufacturing environment conditions from factory production lines. Cases where the emphasis is on data distribution include systems that distribute update programs to distributed embedded systems and systems that distribute digital signage data. Cases where both are done include systems used by automobile companies to collect and distribute road condition data and data related to driving operations (such as areas where the brakes are used frequently).

It is necessary to collect such IoT cases and consider whether you can combine your own technologies and services as analogies, and some companies are probably already doing so. When it comes to developing products and services using IoT as a means, some may argue that the first step is to understand the needs of consumers and customers, and that an approach starting from the seeds is a reversal of subject and object. However, while this is one way of thinking, some are of the opinion that even this is outdated. It is easy to understand from past research and our own experience that when considering customer or consumer insights, there is an aspect in which customers and consumers themselves do not understand their own needs, and therefore an approach starting from the seeds should be seen as an effective method.

As mentioned earlier, advances in machine learning are expected to continue in the future, so we would like to also include utilizing the combination of IoT, big data, and machine learning as business seeds.