CASE 01

Advanced and Automated Vulnerability Management

Automation and Advancement of Vulnerability Information Management for Major IT Vendors

By automating and enhancing vulnerability management, we reduce non-value-added tasks and facilitate the IT organization’s shift toward strategic initiatives, while strengthening cybersecurity resilience through real-time risk visualization.

Challenges

While managing the system operations for dozens of client companies, we found ourselves overwhelmed by day-to-day operational tasks despite having tasks we wanted to tackle to improve service levels for our clients, making it difficult to engage in proactive activities. Furthermore, regarding security risks, manually investigating and collecting vulnerability information and applying it to the systems was extremely time-consuming, and delays in applying these fixes could potentially increase security risks.

Solution

In Microsoft Power Platform, we have implemented a system that automatically imports vulnerability and bug information for major vendors from NIST and vendor websites into a database, and applies the necessary patches to the relevant products in real time.

Benefits of Implementation

This solution has reduced operational tasks related to vulnerability information by approximately 70%. It has also eliminated errors in information gathering caused by human judgment.

CASE 02

Enhancing Management Reporting with Generative AI

Improving Productivity in the Preparation of Management Reports for Chemical Manufacturers

By leveraging generative AI to automate and enhance the processes of collecting, analyzing, and reporting management information, we will accelerate the delivery of information essential for management decision-making, improve the quality of reports—which currently varies by department—and boost the productivity of management control operations.

Challenges

In the budget approval documents prepared for each department on a semi-annual basis, we organize and analyze various data for budget formulation, taking into account information about the chemical manufacturing industry and global trends. However, because the level of detail and quality of the information vary from department to department, it took a significant amount of time (man-hours) to produce reports that satisfied management, which could potentially have impacted management decisions.

Solution

To prepare budget approval materials, we built a system that processes input data in an Amazon Web Services environment (including Amazon Bedrock, AWS Lambda, and Amazon S3) and uses AI to automatically generate the background, rationale, and analysis of the budget in Japanese.
When using AI to generate reports, prompt design is critical to maintaining an appropriate level of detail and minimizing hallucinations. It is also important to standardize prompt specifications and establish them as operational guidelines.

Benefits of Implementation

This solution cut the time required to prepare budget approval documents per department in half. However, it also introduced a new operational task, as the prompts had to be reconfigured each time documents were created. Furthermore, standardizing the level of detail and quality of documents across departments has enabled management to make decisions more quickly than before.

CASE 03

Streamlining Payment Reconciliation with Generative AI

Support for Automating Invoice and Payment Data Reconciliation for Major IT Vendors

In the payment reconciliation process for order management, the workload associated with matching invoices from suppliers with payment data is high, and this work is particularly concentrated at the beginning of each month. To address this challenge, we will provide support for improving operational efficiency through the use of generative AI.

Challenges

While some of the reconciliation of invoices and payment data from suppliers, as well as the task of storing invoice detail data sent via email by various companies on the file server, has traditionally been handled automatically using Excel macros, much of this work for many suppliers is still done manually. Furthermore, the workload is expected to increase due to a planned merger with another company.

Solution

Given the limited time available before our merger with another company, we focused our efforts on leveraging existing tools rather than implementing new ones. As a result, we decided to pursue a strategy of maximizing the use of Microsoft Copilot’s features to improve operational efficiency. Specifically, we provided hands-on support to help users define appropriate prompts for Copilot. In addition, we established a system to document and share this know-how as organizational knowledge so that it could be utilized throughout the company.

Benefits of Implementation

This solution allowed us to nearly fully automate the reconciliation process between invoices and payment data. Although we retained a final step involving manual verification, the overall results were highly effective.
On the other hand, the task of storing invoice detail data sent via email from various companies on the file server could not be fully automated, as it involves complex processes such as password analysis.
However, we were able to achieve partial efficiency gains through a system that analyzes email content to automatically create folders. Furthermore, by documenting the know-how gained from this initiative and applying it to other operations, we were able to create a chain reaction of efficiency improvements.