Interview with Vice President Masayuki Inagaki

Special Feature on Key Measures of Medium-term Business Plan

The most significant change brought about by our promotion of DX has been the shift to data-driven research and development. We set out to prevent valuable information from being scattered throughout the company, and build an infrastructure that we can take maximum advantage of in product development. It is now possible for researchers to enter experimental results into a tablet and immediately share them across the company. We are planning to officially launch this system before the end of fiscal 2023.

In addition, we have rolled out web applications capable of easily performing advanced data analysis to all research laboratories where they are used as tools to support research and development.

At the same time, we have prepared tools for utilizing in-house data that enable ordinary researchers to perform advanced analysis techniques like machine learning on their own for tasks such as search materials data. The needs of society are evolving at an accelerating pace, and it is sometimes necessary to make decisions that run counter to conventional wisdom. In such cases, data becomes the basis for decision-making. We are working to spread an appreciation for the importance of data infrastructure throughout the company, and are training MI personnel who can expand its scope of utilization.

These efforts have already achieved a lot, leading to the development of high-performance materials that surpass the results of previous experiments. These include formulations for materials with high heat dissipation properties, and materials for encapsulation of semiconductors that are flexible yet strong and reliable. Some development sites have achieved significant cost reductions without changing the characteristics of existing products, by replacing materials and optimizing process conditions. Each laboratory is also trying to use MI to reduce the number of experiments that need to be performed. In some cases, we have achieved a four-fold increase in efficiency by utilizing a predictive model for screening experimental conditions, which is quite encouraging.

Today's unpredictable world requires us to rapidly develop products with advanced problem-solving capabilities. Replacing our traditional trial-and-error development style with one that utilizes a scientific approach based on vast amounts of internal and external data is sure to dramatically advance materials development where speed is the decisive factor. We aim to achieve both the advanced functionality and environmental friendliness required of next-generation materials, and improve experimental efficiency during their commercialization, through data-based research and development. We hope this will lead to early acquisition of market share.

We have been promoting digital transformation (DX) in manufacturing around three key pillars: (1) Data collection and storage, (2) Visualization, and (3) Autopilot-based control. Our goal is to build a production system that does not rely on humans. At present, this plan has been implemented at 5 offices in Japan and 5 sites overseas, and is moving forward smoothly.

In production lines, we use various sensors to measure temperature, pressure, flow rate, and vibration to turn this information into digital data. This data is then gathered and made visible through application tools. We carry out quality control for each part number in conjunction with automatic monitoring and autopilot control that notify operators of detected anomalies and the like. We aim to further develop smart factories that do not rely on human workers.

In the manufacturing industry, the declining number of employees due to falling birth rates, and the loss of skills due to seasoned worker retirement, have become issues that can no longer be ignored. We also need to prepare for unforeseen situations such as natural disasters caused by climate change. We want to support societal activity with a manufacturing system that is global in scale, through centralized production information that encompasses all of our overseas sites.

In addition, we plan to actively utilize robotics technology to automate tasks like traverse movement that do not create added value. We have completed the preparatory phase of this project, and expect to implement it within the current fiscal term. Going forward, we plan to roll this out gradually to overseas sites that have become subsidiaries through means such as mergers and acquisitions. We will introduce it as appropriate with consideration for different cultures and values.

With regard to operational reforms, we have positioned RPA as a key method of promoting digital transformation (DX), and have aimed to foster the growth of development personnel while making operations more effective. As of December 2022, there were 260 robots and a workload reduction in excess of 12,000 hours per year. And, in addition to going paperless, we started to introduce AI-OCR^* technology throughout the company during the last fiscal year. By December 2022, we had seen an annual effective reduction of 2,000 hours as a result of automating operations in conjunction with RPA.

The benefits of digitalization go beyond just workload reduction and improved efficiency. I feel that we are creating a virtuous cycle that encourages work-style improvements, such as reducing stress by liberating workers from the excessive tension of working in an environment that does not tolerate mistakes.

In addition, we are now rolling out a no-code tool to the entire company that enables the development of applications without the need for special knowledge or programming skills. This makes it possible for each department to build simple systems without having to rely on the systems department. We are focusing on this system as an effective means of discovering and cultivating DX-capable personnel, and are working to expand its use.

At the same time, we are making efforts to encourage employees to see themselves as the leaders of operational reforms. This entails establishing working groups where employees formulate a vision for each department and review its processes. Then based on the themes which emerge from each working group, they consider for themselves what are the best digital technologies for replacing the operations they wish to make more efficient. We believe it is important for employees to pursue the goals set by the company for improving operational efficiency, while also taking the initiative to make improvements on their own.

We plan to assign data scientists to a wide range of departments within the company beyond just research laboratories, with the goal of promoting data-driven product development. We conducted a one-year-long in-house training program over three terms for a total of more than 50 employees. We intend to establish an internal data scientist certification system by the end of fiscal 2023, and aim to certify 20% of our technical personnel by the end of fiscal 2025.

It is imperative that we secure capable MI personnel in order to handle the remarkable advance of our data-driven society. Since we went into business as a pioneer in plastics, our Group has continued to lead the industry with innovative technologies. We have formed projects led by young employees with the goal of developing personnel with advanced skills and pride in the kind of manufacturing that is unique to Sumitomo Bakelite. These members handle everything from planning to implementation for the training programs conducted at each level. We expect the efforts of young data scientists to be a source of future development. We intend to further increase our presence on the global market, where we have an outstanding market share by building on a solid structure rooted in the wealth of experience possessed by veteran employees.