In the AI industry, more powerful models and larger investment plans are emerging every week.
However, when we turn our attention to the corporate world, the picture is a little different. While the number of companies testing AI in one or two tasks is growing rapidly, This is because there are still not many companies that have established it as a core tool across the entire organization.Wow.
This gap was clearly evident in a Reuters corporate survey released at 8:05 a.m. on August 13 (Korea Standard Time). Among the Japanese companies that responded to the survey, only 16% reported using AI as a core tool across the entire organization. The remaining 84% either use it only for certain tasks, have not yet decided whether to adopt it, or are not even considering adoption at this stage.
| Only 16% use AI Key Tools for WarriorsUsed as
This survey, conducted by Nikkei Research on behalf of Reuters, took place from July 29 to August 6, 2026. Questionnaires were sent to 510 Japanese companies, of which 219 responded anonymously.
Therefore, rather than viewing the 84% figure as a definitive value that applies to all Japanese companies, it is more accurate to understand it as a trend reflected by the companies that responded to this survey.
The response results can be broken down into the following stages:
|
Stages of AI Adoption |
Response Rate |
|
Used as a key tool across the entire organization |
16% |
|
Used in certain departments or for specific tasks |
About 60% |
|
A decision on whether to implement it has not yet been made. |
18% |
|
Not considering implementation |
6% |
The largest group consisted of companies that use AI only in certain departments or for specific tasks. Rather than rejecting AI outright, they are testing it out with tasks that are easy to start with—such as document drafting or information organization—but have not yet expanded its use to the company-wide workflow.
Not all companies that adopted AI across the board utilized it to the same extent. One respondent explained that while the company uses AI company-wide, its actual applications are limited to tasks such as document creation. Therefore, even the 16% figure does not necessarily mean that all core business operations have shifted to being AI-driven.
| This does not mean that 84% do not use AI at all
The aspect that deserves the most attention in this survey is 84%That's what it means.
That 84% doesn't represent the percentage of companies that don't use AI at all. About 60% are already using AI in certain departments or for limited tasks. The 84% figure is the sum of all other responses, excluding the 16% that use AI as a core tool across the entire company.
According to the 2026 Information and Communications White Paper released by Japan’s Ministry of Internal Affairs and Communications, 86.4% of Japanese companies surveyed reported using generative AI in at least one business process. In contrast, only 16% of respondents to a Reuters survey said they use it as a core tool across the entire company.
Since the two figures differ in terms of the survey population and the way the questions were asked, they cannot be compared directly. However, Data showing that there is a significant difference between companies that have used it at least once and those that have integrated AI into their company-wide operationsIt does have meaning.
If you look solely at AI adoption rates, it may seem as though AI has become quite widespread, but the reality is that far fewer companies have actually changed their work processes and decision-making structures as a result.
| Companies aren't cutting their AI budgets, but 31% haven't decided on a direction yet
A similar trend was evident in the AI budget outlook for the next one to two years.
|
Annual AI Budget Forecast |
Response Rate |
|
An increase of 50% or more |
4% |
|
An increase of 10% or more but less than 50% |
21% |
|
Single-digit increase |
30% |
|
Generally unchanged |
14% |
|
Haven't decided yet |
31% |
|
Decrease |
0% |
55% of the responding companies expected their AI budgets to increase, and not a single company said it would cut its budget. This means that companies haven't given up on the potential of AI itself.Wow.
However, it’s also important to note that 31% have not yet determined the scale of their investment. Which tasks should AI be applied to in order to generate actual results, and what are the costs associated with security and data management? A sign that quite a few companies have not yet found an answer to how to measure return on investmentIt can be seen as...
Ultimately, the question we're facing now is less about whether to adopt AI and more about where and how much to invest.
| Why It's Difficult for Pilot Programs in Some Departments to Be Rolled Out Company-Wide
Simply purchasing AI tools and providing them to employees isn't enough to ensure company-wide adoption.
Each department uses different data formats and work procedures, and access permissions for sensitive information also vary. We also need to determine who will review and take responsibility for the results generated by AI, as well as the procedures for correcting errors when they occur.
This also requires integration with existing systems, employee training, security standards, and performance measurement methods. Without this foundation, AI is likely to remain nothing more than a personal assistant tool for drafting documents or summarizing meetings.
In this survey as well, an official from a real estate company responded that they weren't sure where exactly to apply AI. This means that while the environment for using the technology has been established, the business problems to be solved and the methods for implementing it have not been sufficiently defined.It can be read as.
| What's needed is a redesign of the workflow structure, not a more powerful model
For a company to advance its use of AI, it needs to streamline its workflows before introducing additional, state-of-the-art models.
Where do repetitive tasks occur, what data is required, and which points must be reviewed by a person?We need to start by deciding that. After that, we should identify results—such as time savings or fewer errors—from small tasks and then expand the proven methods to other departments.
It's not enough to just track the number of times AI is used. How much have existing business metrics—such as processing time, cost, quality, and customer response time—actually changed?We need to look at these together. That's the only way to turn simple usage into improved organizational productivity.
In other words, the key to enterprise AI isn’t about providing the same chatbot to every employee. It’s about designing the data, permissions, review procedures, and accountability structure together so that the system can operate reliably even after the AI is implemented.
| It is difficult to view this as a problem unique to Japan
Although this survey targeted Japanese companies, the questions themselves can be applied directly to companies in other countries as well.
What matters more than how many employees are using AI is whether AI is integrated into actual work processes. There’s a world of difference between simply reducing the time it takes to draft documents and transforming core processes such as product development, customer service, supply chain management, and decision-making.
Going forward, the AI gap between companies is likely to stem more from the depth of application than from simple model performance. This is because companies that have streamlined their business data and procedures—even when using the same AI—can quickly achieve results, whereas companies that leave it solely to individual employees to experiment find it difficult to expand the scope of its use.Wow.
| The key isn't the adoption rate, but whether work practices have actually changed.
The 16% and 84% figures from the Reuters survey do not signal that the spread of AI has stopped. Rather, they suggest that many companies have moved beyond the initial phase of testing AI and have entered the more challenging stage of enterprise-wide adoption.
The fact that no company plans to cut its budget also suggests that AI investment is likely to continue. However, the next round of competition will hinge not on who purchases the most AI tools, but on who can more specifically redesign their workflows to align with AI.
The true adoption rate in the AI era isn't measured by the number of accounts, but rather by Judging by the actual number of changes in work tasksI have to.
Thank you for reading!
References
#AI #Generative AI #Japanese Companies #Enterprise AI #Warrior AI #AI Adoption #AX #Digital Transformation #IT News
