It will soon be four years since ChatGPT was released and the use of AI began to be called out. During this time, we have repeatedly encountered the question of "who will wave the flag" when it comes to using AI within our company. If you hand out tools, people in the field will find ways to use them, and success stories will spread laterally. That bottom-up picture was often talked about in the early days of its introduction.
But now, that composition is losing its shape. In a survey of 400 people conducted in May by marketing support company Thrista (Yokohama City), the AI usage rate for section managers was 67.3%, while it was only 25.4% for general employees. More than 50% of employees are in the management and general manager ranks, so it appears that the bottom-up structure that was expected has not materialized.
There are various reasons for this, but I believe that one of them is a structural problem that makes it difficult to successfully utilize AI from the bottom up. Field-led AI utilization is not successful - why does this situation occur?
Non-regular employment and work style reform, two factors that impede bottom-up efforts
We believe that there are two structural characteristics that Japanese companies face.
The first is the spread of non-regular employment. The use of non-regular employment has been in full swing for about 30 years, but in order to hand over tasks to non-regular employees, it is necessary to subdivide tasks and establish procedures.
As a result of operating in this state for many years, the unit of business itself has become fixed. Changing your workflow means redrawing the boundaries between dozens of fragmented tasks, and the cost of change is higher than you might think.
Another is work style reform. With overtime being considered a bad thing, workplaces have lost the leeway to be creative with their workflows. Although this is a slightly older survey, 60% of respondents cited ``lack of emotional leeway'' as the reason for the drop in satisfaction due to work style reform, and 45% cited ``decreased productivity.''
Field-based business improvement, including AI, is an activity that takes place "outside" of routine work. It is unreasonable to assume that improvements that originate from the workplace will accumulate in a situation where the external margin is being eroded by the systemic problems of the expansion of non-regular employment and work style reform.
As an example, let me introduce an actual episode. When we tried to improve operational efficiency at an IT company, we encountered strong resistance from a team made up of contract and temporary employees. The object of resistance is not automation itself. This was a verification work that compared the results of automated processing and conventional manual processing.
For your team, the transition period will be a double-duty process, with one manual task and one validation step at a time. The team clearly opposed the idea of having to work overtime for that amount of time. If you think about it, it makes sense; for them, automation means that their responsibilities will disappear, so there is no reason for them to take on temporary overtime.
``I'll do my best now to make things easier'' is a common trait among IT engineers, but it's hard to accept for other professions. What was needed at this time was not to persuade the on-site staff, but to decide who would pay for the man-hours during the verification period. It is outside the discretion of the field.
The main premise is that regardless of the industry, whether it's an IT company or manufacturing industry, the workplace will resist changing the established workflow. However, if executives dismiss this resistance as ``maintenance on the job'', they may be making the wrong move.
