Regarding this, IBM Japan's Yuko Kawakami (Managing Executive Officer, General Manager of Growth Strategy Division), who took the stage after Mr. Hayakawa, explained about ``collaboration between people and digital workers'' while introducing IBM's own use cases. By the way, digital workers are AI agents. The company refers to them as digital workers when explaining their use cases, so we will refer to them as such in this article.
"Today, our colleagues at IBM are not just humans; we have more than 4,000 digital workers working together on more than 450 project teams. Digital workers are filling jobs traditionally held by humans, including consultants, analysts, architects, engineers, project managers, designers, and creators."
Mr. Kawakami explained this while showing Figure 2.
So what are people doing? He listed four areas: ``determining policy,'' ``taking responsibility,'' ``ethical judgment,'' and ``handling exceptions.'' He added, ``This initiative is not just about introducing AI tools.The composition of the team has changed, and the members are no longer just humans.This is the current way IBM works.''
Figure 3 shows how IBM operates digital workers as "human resources." The company employs dozens of job-specific digital workers who earn the same skills certification badges as employees and are deployed to their respective project teams. According to Mr. Kawakami, ``The same system as human resources is applied to digital workers: recruitment, certification, deployment, and retirement when the mission is completed.''
Why IBM promotes “collaboration between people and digital workers”
IBM has been promoting this kind of "collaboration between people and digital workers" as an initiative for Client Zero (zero customer) since 2023, and so far, "We have broken down the entire company into 490 business flows, redesigned 70 of these processes based on AI, and created a productivity improvement effect of 4.5 billion dollars (as of 2025)" (Mr. Kawakami). As a result, the company has created more than 150 AI use cases across a wide range of operations, and the funds generated are being reinvested in focused areas (Figure 4).
Mr. Kawakami also explained the conditions for achieving results from the use of AI based on the company's research, while showing Figure 5.
On the left of Figure 5, the vertical axis shows ``Did you just apply AI or redesign the job?'' and the horizontal axis shows ``Did you just try it out in some operations or did you roll it out across the company?'' As a result of the survey, only 15% of companies achieved results by redesigning work and rolling it out company-wide. However, it was also found that these 15% of companies outperformed their peers by 73% in revenue and 11% in operating profit margin. Mr. Kawakami said, ``The AI disparity is also reflected in these survey results.''
Why is the use of AI in many companies limited to the results of individual operations and not expanded to management results? ``This is because there is a lack of a foundation for linking AI to business results,'' said Mr. Kawakami, citing the following four issues as obstacles to the company-wide rollout of AI results.
So, what is needed to resolve these four issues? Mr. Kawakami stated, ``Each of these issues requires not only business redesign (operational level), but also a corporate management system that continues to redesign and expand its implementation (foundational level),'' and showed the responses at the operational level and foundational level for each of the four challenges (Figure 7).
He added, ``I would like people to be careful because the most likely stumbling block when attempting to collaborate between people and digital workers is the content at the foundation level.''
For more information on IBM's new solution ``IBM Enterprise Advantage,'' which supports these efforts in companies, please see the presentation materials (translated from IBM's May announcement).
Collaborate with digital workers instead of using AI tools
Finally, a word from the author. I felt that the ``collaboration between people and digital workers as IBM's own use case,'' which I heard at this IBM Japan press conference, was an indication of the future of corporate management. What I would like to emphasize once again is that ``rather than using AI tools, we should collaborate with digital workers.''
Furthermore, AI disparity can occur not only in companies but also in individuals. In order to solve this problem, it will be essential to promote "AI for social infrastructure." Consistency will also be required in the relationship between AI for social infrastructure and AI for corporate management infrastructure. We will continue to monitor these developments from this perspective.
