Customer Interviews

Thank you for taking time out of your busy schedule to give us this interview today. I would like to start by asking you about your company’s AI initiatives. How do you position AI in your company as you strengthen R&D?

We divide our efforts into two main categories: “AI for business use” and “AI for R&D use. For business, we are trying to use AI in the electric power business to make various predictions as part of the “SWCC Smart Stream Business. For example, we are trying to predict whether or not there will be accidents at construction sites. We are building a system that uses AI to predict areas and factors that may cause accidents in the future, based on data from past workplace accidents within the company, and then alerts people to the possibility of such accidents. This AI technology is now being used for commercialization. In research and development, we are currently working on informatics like the one we are asking your company to do, and on the creation of a database of reports using generative AI and other technologies. Since we possess the fundamental technology, we are currently examining the possibility of applying AI to various situations.

Thank you very much. In your company’s mid-term plan, you have set data-driven business and solution as the business models to aim for. Could you explain what these involve?

Yes, we advocate a “data-driven” business model, using accumulated data to speed up the prototyping process, etc. We are trying to create a cycle in which prototypes are made and evaluated by making full use of AI and simulations. In the future, we are considering the possibility of turning this into a solutions business. It is still a long way off, but we would like to launch it as a business by 2030 or so.

I see. That is a wonderful initiative.

Now, let me ask you about Multi-Sigma. Please tell us about the challenges you faced when you started using Multi-Sigma and what motivated you to start using it.

At that time, even without in-depth knowledge of materials informatics, predictive applications were gradually emerging as long as we could properly prepare the explanatory variables. Our company was already using one of them. However, due to the “AI black box problem,” we still felt that we didn’t truly understand the inner workings, even though we were told we were using such models. When we actually conducted analyses, we faced challenges like “not getting the expected results,” “how should we train the data we’ve collected,” and “what can we do to obtain the appropriate answers we’re seeking.” That’s when we discovered Multi-Sigma. While Multi-Sigma is built very simply, it allows us to select from several AI models generated through its auto-tuning feature. I found it quite unique that we could set up various patterns of explanatory variables, build highly accurate models, and compare and evaluate each one. With this capability, AI was no longer a black box; it was significant that we could perform analyses with the conviction that “we are obtaining these results precisely because the model is highly accurate.” As we continue to use Multi-Sigma, I feel that our team members can gradually deepen their knowledge of informatics.

What are your expectations after using Multi-Sigma?

As for expectations, I hope that our employees will become true experts in their fields. I hope that by using Multi-Sigma, they will be able to refine their sense as professionals throughout this process. Currently, we have not reached that point yet. However, I feel that the causal relationship within the “black box” of our process is gradually becoming clearer as people are actually involved in process informatics and materials informatics through Multi-Sigma. I anticipate that the level of our manufacturing personnel will rise as informatics reveals aspects that were previously attributed to craftsmanship, and as our employees sharpen their skills and senses.

Could you comment on the expectations and outcomes of our consulting services?

When we introduce informatics software, we often sign a consulting contract to receive professional advice. However, among all the consulting we’ve experienced so far, Dr. Kawajiri has been the best. Perhaps it was because we had been working with Dr. Kawajiri for some time, he gave us the most direct advice on the issues we were struggling with and what steps we should take. We are very grateful for his support. We are currently engaged in manufacturing, and there are still areas where we rely on the skill of a master craftsman. By applying informatics to these areas, there is room to develop more efficient, stable, and high-performance products, and I believe we need to leverage the best aspects of informatics to make this possible. looking further into the future, I would like to transform informatics itself into a business. I would be very happy if we could continue to collaborate with AIZOTH, benefiting from their experience and advice, as we further develop informatics and turn it into a viable business model.

I am very honored. Thank you very much for your valuable feedback today. We will incorporate your feedback into our future company services. We look forward to working with you in the future.