FRONTEO Inc. (Headquarters: Minato-ku, Tokyo, President: Masahiro Morimoto) used FRONTEO's artificial intelligence KIBIT, which is the third project of the Fintech demonstration experiment hub conducted by the Financial Institution in May 2018. For the first time since the results of the demonstration experiment on improving the efficiency of compliance check operations of financial institutions were announced, the subsequent efforts of each company that participated in the demonstration experiment were discussed at the "5st Japan International Financial System Forum" held on February 3, 2020. Introduced in "2".

This demonstration experiment was conducted with the cooperation of The Bank of Tokyo-Mitsubishi UFJ, Ltd., Resona Bank, Limited, Bank of Yokohama, Ltd., and SMBC Nikko Securities Co., Ltd. as participating financial institutions, and there is a suspicion of suspicion of non-compliance at financial institutions. This is a comparison of business productivity between "checks by people only" and "checks using KIBIT" for the work of finding data.How many "correct answers = records to be found in the check work" can be found within a fixed time for checking whether there is any compliance violation in the explanation contents of the staff and extracting complaints from customers included in the reception record etc. Was quantitatively compared and measured.As a result of the measurement, the correct answer detection rate when using KIBIT is equal to or higher than that when only humans are used, proving that KIBIT was able to fully learn the tacit knowledge of humans, and AI is used in compliance check work. It was found that this can be expected to improve accuracy and efficiency.

In monitoring work such as compliance check of financial institutions, many people had to gather and review with the conventional method, but by using KIBIT, it is possible to replace the work by humans, and work efficiency We will realize the improvement of business quality.In addition, since it is not necessary for many people to gather, it is an epoch-making solution that leads to anti-virus measures, which has become a major issue in Japan recently.

Since then, the financial institutions participating in the experiment have been promoting actual operation.In that process, I found that there were business issues before the actual operation.For example, there are a huge number of compliance perspectives to be confirmed, and there are various ways to keep reception records and the management system is not unified.In addition, the issues regarding how to use KIBIT have been clarified. For example, when describing multiple products, it may not be possible to determine which product the specific description is for.While searching for ways to solve these issues, participating financial institutions are currently working to start actual operations in fiscal 2020 with the aim of improving and streamlining the monitoring operations at the time of product sales and complaint monitoring operations.

FRONTEO and participating financial institutions will continue to utilize and apply this demonstration experiment in actual business with the aim of achieving both improvement of business productivity and realization of "customer-oriented business operation (fiduciary duty)". I will continue to explore that.

■ About KIBIT
"KIBIT" gives a document selected based on "tacit knowledge" such as past examples and the intuition and feeling of experienced people as "teacher data", and learns the characteristics of the document, along the judgment axis. It is an AI engine that efficiently extracts the documents you want to find. "KIBIT" is characterized by analyzing and sorting a large amount of data in a short time with lightweight operation even if the amount of teacher data is small.

FRONTEOabout　URL: https://www.fronteo.com/

FRONTEO Inc. has developed its own artificial intelligence engine "KIBIT".^®(Kibit) ”and“ concept encoder^®(Concept encoder) ”is a data analysis company that supports information analysis such as big data.Established in August 2003 as a company that supports e-discovery (electronic discovery), which preserves, investigates, and analyzes electronic data necessary for international litigation, and digital forensic investigations.In-house developed data analysis platform "Lit i View"^®(Lit Eye View) ”,“ Predictive Coding Predictive Coding^®(Predictive Coding) ”technology is used to provide legal proceedings support to companies.The unique artificial intelligence-related technology cultivated and developed in this legal business learns "tacit knowledge" such as expert experience and intuition, and realizes prediction of future behavior from analysis of human thoughts.We have expanded into areas such as life science and business intelligence, and have achieved results in "work style reform" in addition to FinTech and RegTech. Listed on TSE Mothers on June 2007, 6.The capital is 26 thousand yen (as of March 2,559,206, 2019).