Establishment of a New AI Promotion Department

Over the past six months, ChatGPT and other generative AI tools have been attracting much attention, and the nature of business operations, especially in the IT industry, is undergoing a major change. A while ago, at Mitsue-Links, we established an Automation Department that was tasked improving business efficiencies through the use of RPA, however, we recently rebranded this department as the AI Promotion Department, with the aim of researching and promoting not only automation but also the use of AI. With recent graduate recruits joining the department, we are now making efforts with a fresh mindset and seeking ways to utilize technologies that have never been used before. Based on a vision of "using AI to develop productivity-enhancing solutions that cannot be achieved through conventional methods, thereby increasing our competitiveness in the marketplace," we intend to create technologies to improve productivity more broadly than ever before.

I have worked with a variety of automation technologies, and, going forward, I predict that with the use of AI, automation will become even more accessible to many more people. Firstly, the bar for coding for automation has been lowered dramatically. I've heard from many coders that they've found using generative AI as a coding aid has improved their coding efficiency. Furthermore, even those who don't usually code can now simply ask AI to write basic script. With the advent of AI many people having their first experiences of code creation, now anyone can easily use the tools to improve work efficiency.

Also, I feel that the use of AI has made it simpler to extract tangible outputs from relatively vague, obscure ideas. For example, if you inform ChatGPT, that you "want to write a blog about AI applications, and need 10 title ideas," the tool will immediately propose answers. For philosophies of visual content to be used in a blog, writers may consider using image-generating AI tools. Having tangible imagery can further support thinking and make it easier to share ideas with others. With a little ingenuity, a range of output can be created – while this is useful for writing code and creating text content, it can also improve thinking and communication efficiencies.

On the other hand, however, AI generated output is not always appropriate. For example, for generative AI, code is one area with a fairly high accuracy, even so, however, output sometimes doesn't result in the initially intended script. Also, rather than merely using the output of AI as is, it's likely that it will be deployed after appropriate adjustments are made. It is my opinion that even if coding itself becomes less time-consuming, a proper understanding of the code (i.e., the ability to examine the AI output) will become increasingly important.

The AI Promotion Department, while deliberating on the attributes of various technologies, is currently researching effective ways to utilize AI in business operations. Specifically, we are examining both the possibility of providing client-oriented services to enhance work efficiency and further improving quality in internal operations. Regarding our internal use, we've already begun to see positive effects from the introduction of AI-related technologies on a trial basis in some operations that handle routine data, such as file list creation. Although this is a rapidly evolving field, through closely monitoring the latest trends and skillfully working with AI tools, we aim to develop effective solutions.