(4) Preserved emails and chats of several Company D subjects and conducted digital forensics. KIBIT conducted up to the first review of the preserved and restored data.
In this case, KIBIT contributed to (1) precise targeting of review targets, (2) early detection of important/relevant documents, and (3) optimization of human, time, and monetary costs through cutoff.
The number of review targets was approximately 2,100,000, but the search, de-duplication, and email threading process narrowed the number of review targets to approximately 450,000.
After the KIBIT score ranking, we found about 6,800 documents, or 84% of all important/relevant documents, after reviewing about 45,000 documents (10% of the total) on the first day.
The human review was terminated when approximately 150,000 documents were reviewed, starting with the top 3) documents with the highest scores. The remaining 300,000 documents were reviewed only by KIBIT. By cutting off the human review, the human, time, and monetary costs that were initially expected to be incurred were eliminated, contributing greatly to cost optimization. In addition, since the Elusion Test (human review of approximately 1,600 documents randomly sampled from the cut-off 300,000 documents) was conducted on the cut-off 300,000 documents, and it was confirmed that no important documents related to the inappropriate conduct in question were included in the cut-off 300,000 documents, the first review was conducted to confirm that no important documents related to the inappropriate conduct in question were included in the cut-off 300,000 documents. Therefore, the first review covered approximately 97% of the documents related to the inappropriate conduct in question, based on a proportional point estimate.
Finally, approximately 300,000 cases, or 67% of all cases covered, were reviewed only by KIBIT (AI Only Review), which contributed to shortening the turnaround time and optimizing the monetary cost by more than 60 million yen by not conducting human review.
Outside Investigation Committee Company D (Insurance Industry)
Violation of the Antimonopoly Law
Investigation to confirm facts of improper conduct
Point
-
Investigation terminal
E-mail, chat
-
Scope of investigation
Up to 1st review
-
Contributions of using KIBIT
(1) Accurate narrowing of review targets
(2) Early detection of important/relevant documents
(3) Optimization of human, time, and monetary costs through cut-off -
Effects of using KIBIT
(1) Narrowing down the number of documents to approximately 450,000
(2) Detection of 6,800 documents (84% of all important/relevant documents) after reviewing about 45,000 documents (10% of all documents) on the first day
(iii) Shortened turnaround time and optimized monetary costs by more than 60 million yen
Examples of Investigations by Third-Party Committees
Some of the forensic investigations FRONTEO has conducted in the past for third-party committees
