According to a survey conducted by TWOSTONE&Sons, Inc. targeting 111 executives and managers driving AI/DX in companies with 300 or more employees, it was revealed that 91.9% of in-house AI projects have experienced stagnation or failure.
Over 9 of In-house AI Projects Stagnate or Fail, Driven by Issues in "Strategic and Conceptual Design"
PLUS ULTRA by Amenoyomi
This article is a translation. Read the Japanese original
The survey results showed that the most common stage where projects tend to stall is the "stage of deciding the purpose and theme of AI utilization," at 57.7%. Additionally, regarding factors important for promoting AI utilization, "strategic and conceptual design of AI utilization" reached 58.6%, surpassing the selection of technology and tools (37.8%).
Regarding the background of project stagnation and failure, the most frequent response was a lack of "the ability to link management and business issues to AI utilization" at 58.8%. As for the impact caused by stagnation, the reduction or freezing of projects and investments was the most common at 56.9%, followed by delays in full-scale implementation or company-wide deployment at 55.9%.
Furthermore, regarding areas that are difficult to address solely within the company, "strategic and conceptual design of AI utilization" was the most common at 58.6%. As a reason for this, 53.6% of respondents stated that there is a "lack of personnel who can understand both management, business, operations, and AI technology."
PLUS ULTRAby Amenoyomi
The stagnation of in-house AI development is not simply a matter of failing to find the right tools; the primary cause is stumbling during the "upstream design" phase, which involves defining the purpose and theme. The survey shows a structure where challenges are concentrated in the upstream processes of strategic and conceptual design rather than the selection of technology and tools, with 57.7% of respondents feeling stuck at the stage of deciding the purpose and theme of AI utilization.
The reason for stumbling at this design stage is the absence of personnel who can understand both management, business, and operational tasks, as well as AI technology, and can concretely connect them. As a reason why it is difficult to handle these issues within the company alone, "lack of personnel who can understand both management, business, operations, and AI technology" was cited most frequently at 53.6%. This indicates that the bottleneck is not merely technical ability, but the ability to translate business issues into AI specifications.
These deficiencies in strategic design lead to more than just schedule delays. As a result of proceeding with insufficient design, the costs and man-hours invested do not lead to results, leading to cases where projects and investments are reduced or frozen in 56.9% of instances. Furthermore, this directly leads to organizational risks, such as increased distrust or resistance toward AI utilization at the operational level due to the failure to achieve expected results.
Sources
- AI内製化、9割超が「停滞、頓挫」を経験 つまずく原因は“技術”ではなかった （ITmedia AI+、2026-09-25）
- TWOSTONE&Sons