One HR confirms business reduction effects as a result of joint research with Osaka City Waterworks Bureau on creating personnel placement proposals using HR data and AI

One HR Corporation (Headquarters: Shinagawa-ku, Tokyo; Representative Director and President: Yuzaburo Karasawa; hereinafter: our company) confirmed, as a result of the "Joint Research on Creating Personnel Placement Proposals Utilizing HR Data and AI" (Note) conducted with the Osaka City Waterworks Bureau, which began in January 2025, that it is possible to extract candidates to a certain extent by matching human resource information with position requirements using generative AI (LLM). It also confirmed challenges for practical implementation and obtained results indicating that a reduction in operations related to candidate extraction and review can be expected in the future.

Note: For details on the "Joint Research on Creating Personnel Placement Proposals Utilizing HR Data and AI" conducted with the Osaka City Waterworks Bureau, please refer to our press release: "One HR Concludes Agreement with Osaka City Waterworks Bureau on Joint Research for Creating Personnel Placement Proposals Utilizing HR Data and AI."

Verification Points

To more effectively create personnel placement proposals, which the Osaka City Waterworks Bureau has traditionally carried out, by utilizing human resource information, the following two points were specifically verified:

・ Verification of whether more effective and efficient personnel placement proposals can be created by utilizing AI.

・ Verification of what kind of data collection, accumulation, and organization is necessary for AI to create personnel placement proposals.

Implementation Details

The personnel placement proposal creation work at the Osaka City Waterworks Bureau involves:

・ Successor placement for managerial positions (section chief, deputy section chief, unit chief) that become vacant due to transfers, etc.

・ Relocation of staff members who have gained experience in the same department for a certain period to departments where they have no experience, in order to acquire various work experiences.

These two types of placement considerations are combined and carried out. For each, we conducted efficiency verification utilizing generative AI (LLM) and combination optimization algorithms.

＜Model Diagram for Successor Candidate Determination for Managerial Positions＞

＜Model Diagram for Generating Staff Relocation Proposals＞

Research Results

・ In determining successor candidates for managerial positions, it was confirmed that generative AI (LLM) can generate concise and natural Japanese person summaries and, by comparing these person summaries with simple position requirements (set by the Osaka City Waterworks Bureau at around 100 characters), it is possible to recommend successor candidates and generate explanatory texts for their rationale to a certain extent.

・ As a challenge, the necessity to standardize expressions was confirmed regarding the impact of variations in expression within the generated explanatory texts on matching.

・ Regarding the above challenge, if accuracy and reliability for practical operational levels can be improved, it is estimated that approximately 150 hours of effort can be reduced in the current managerial-level placement proposal review work, which currently requires about 170 hours for tasks such as searching various HR-related documents and confirming their contents for candidate extraction for each position.

・ It is believed that position suitability can be further enhanced by describing position requirements more specifically and increasing the clues for generative AI (LLM) to extract successor candidates.

・ For generating staff relocation proposals, an automatic generation logic was implemented using simple conditions such as the individual's wishes and the supervisor's recommendations. However, by incorporating requirements as a collective value of the members constituting the organization, such as personnel composition balance by grade and training conditions like career plans, into the logic, the generation of more realistic placement proposals can be expected.

What is "One HR"?

Surpassing 600,000 paid users! A human resources and labor management system that provides one-stop solutions for operational efficiency in labor management, attendance management, payroll calculation, talent management, and workflows, enabling integrated management of human resource information. It flexibly adapts to the growth of organizations, from general companies to public institutions such as government agencies and local governments, fully supporting employee growth and operational efficiency.

・Service HP: https://onehr.jp/

・Corporate HP: https://onehr.co.jp/

・Official X: https://twitter.com/onehr_jp

・Official Facebook: https://www.facebook.com/onehrjp

＜Contact for this release＞

One HR Corporation, Public Relations E-mail：pr@onehr.jp