A case study of SHIFT’s use of AI collaboration capability in the hiring process has been published. It presents facts and data on how collaboration between AI and people is evaluated, highlighting the shift toward hiring evaluations that emphasize thinking processes as well as results.
株式会社SHIFT published a case study on visualizing engineers’ ability to use AI based on facts and data rather than intuition when hiring engineers in the AI era. The implementation used HireRoo’s evaluation function and shifted the hiring criteria from an emphasis on results to an emphasis on thinking processes.
One of the three key points of the implementation was the establishment of a system for supporting interviewers’ instincts with data, making it possible to evaluate candidates based on a definition of AI collaboration capability. The second was shifting technical evaluation beyond coding to the thinking process, establishing a system for examining how candidates used AI and how they gave instructions based on assessment data. The third pointed out that whether verification was performed affected rejection decisions, indicating a bias toward evaluations that emphasized verification.
The case study includes a summary of SEI evaluation data, including more than 1.17 million evaluation data points as of July 2026, as well as the average scores for the 5 areas of AI collaboration capability. The lowest-scoring area was collaborative reasoning at 48.6%, followed by environment setup at 57.6% and output quality at 61.3%. The study pointed to a lack of ability to verify one’s own work while interacting with AI.
The announcement also covered the release of a white paper and the planned broadcast of SHIFT’s taxi commercial. The white paper includes an overall framework for AI collaboration capability and an explanation of the evaluation data. The taxi commercial is scheduled to air for 2 weeks from 14 September in GO-equipped vehicles in the Tokyo area.