A survey of generative AI use among small and medium-sized enterprises found that while individual tasks are becoming faster, many respondents said improvements in overall team speed remained limited.
According to the “Survey of Generative AI Use in the Workplace by Small and Medium-Sized Enterprises,” released by TANKA Japan on 22 October 2026, 64.2% of the 500 respondents, who work for domestic small and medium-sized enterprises with fewer than 100 employees and use generative AI in their work, said their own work had become faster. By contrast, only 51.8% said the speed of work across their team had improved.
A total of 60.0% said they felt that “my own work has become faster, but the team as a whole has not become much faster.” The biggest causes of delays were waiting for replies, checks and approvals from others, at 28.2%, and searching for necessary information and the background to past decisions, at 23.6%. Work arising between people accounted for a combined 71.0%.
The survey also highlighted the limited sharing of AI-generated results. Only 24.0% said their teams shared documents created with generative AI or information organized using it, while 58.6% said work came to a halt when the person responsible was absent or left the company. In addition, 39.2% said their AI knew little about the circumstances of their team.
The functions respondents most wanted from AI in the future were tools to facilitate checks, handovers and information searches between people, at 42.0%, and tools to further speed up their own work, such as document preparation, at 41.2%. A total of 76.2% said they would like to try AI that understands their team’s history and responsibilities and uses that knowledge to move work forward.