Currently, technological innovations of the fourth industrial revolution, including AI, are emerging one after another. However, the market created by these technological innovations has not yet been formed, and the number of companies that have entered the market has not yet emerged. It is certain that new business opportunities exist, but there is no objective data to analyze them.
Something similar to this happened when the Internet was commercialized 30 years ago. While Japanese and European (especially Japanese) companies were stuck in such a situation, U.S. companies such as Google and Amazon pushed strongly into the unexplored frontier and established a dominant position. There is no longer time to hesitate if you want your business to ride the next wave of technological innovation.
The “Concept Reserch Method” is a technique for exploring business opportunities related to emerging technologies.
Around 1996-1997, Kunihiro TADA argued that business research should include not only factual research, but also conceptual research, for the sake of large Japanese companies who were not aware that the Internet was changing the world. However, the use of Husserl’s phenomenology to theorize Cocept Research made it increasingly difficult for business people to understand. Reflecting on the importance of presenting technical tools rather than philosophical arguments, Tada became involved with Kohonen’s Self-Organizing Map (SOM) in 1998.
SOM is precisely because it is a technical tool that generates concepts from data. In 2000, we translated Viscovery SOMine into Japanese and began selling it in Japan. It is the most complete SOM data mining system on the market. Then, in 2003, we began selling Hugin, a Bayesian belief network (BBN). Hugin is a pionier product which has contributed greatly to the practical application of BBN.
By combining SOM and BBN, a computer-assisted KJ method can be constructed. The KJ method is a method of analyzing qualitative information invented in Japan in the 1960s. KJ are the initials of the originator, Jiro Kawakita, a cultural anthropologist. A similar approach is known internationally as the Grounded Theory Approach (GTA). These methods have the following steps in common:
- Field research collects many pieces of information.
- Classify pieces of information by similarity. (The KJ method recommends repeating this classification over and over again “until the scales fall from your eyes.”)
- Extract common characteristics within groups created by classification.
- Explain relationships between groups or characteristics. (The KJ method creates a diagram.)
We can achieve 2 and 3 using SOM and statistical tests, and 4 using BBN. However, handling qualitative information requires advanced text mining techniques. Unfortunately, traditional text mining is not powerful enough to do this. OpenAI’s Chat-GPT was exactly the solution for this. Initially, Chat-GPT was capable of use words like a human, but the content was not very reliable. However, after about a year of testing, it has become clear that it can extract document features using more sophisticated concepts.
Thus, we are now able to use OpenAI techniques to extract features from documents, analyze concepts with SOM, and model relationships between elements with BBN.