Persol Business Process Design Conducts "Demonstration Experiment on Knowledge Generation Utilizing Call Recording Data from Telephone Inquiries" in Setagaya City

Persol Business Process Design Co., Ltd. (Headquarters: Minato-ku, Tokyo; Representative Director and President: Kazuyuki Ichimura; hereinafter "Persol Business Process Design"), a member of the Persol Group, which upholds the vision of "Work and Smile," has completed a "Demonstration Experiment on Knowledge Generation Utilizing Call Recording Data from Telephone Inquiries" in Setagaya City, Tokyo (Mayor: Nobuto Hosaka). In this demonstration experiment, call recording data from inquiries received from citizens at the main business office telephone were collected and analyzed. This was done to visualize the response know-how based on staff experience and judgment for accurate referrals, and to structure knowledge with an eye towards automated responses by AI.

Click here for details on the "Contact Center Consulting" service:

https://www.persol-bd.co.jp/service/contactcenter/s-cc/service/consulting/

■Background

In recent years, with the diversification of citizens' lives and administrative services, the content of inquiries to local governments has become more complex. Setagaya City provides contact information for each department through its website and other channels, but in some cases, inquiries concentrated on the main business office telephone because callers did not know where to direct their questions. At the main business office telephone, staff handle inquiries in parallel with their regular duties, and answering calls during their daily work posed challenges in terms of response burden and operational efficiency.

Furthermore, since a wide variety of inquiries are received, accurately understanding the purpose of the call within a limited time and smoothly connecting it to the appropriate department has relied on the experience and knowledge of each individual staff member. This has led to issues such as response skills being dependent on individual staff experience, becoming tacit knowledge, and not being sufficiently accumulated as data.

Against this backdrop, with the aim of building a system that can deliver information more clearly and accurately without keeping citizens waiting, the Setagaya City DX Promotion Department and Persol Business Process Design collaborated on a "Demonstration Experiment on Knowledge Generation Utilizing Call Recording Data from Telephone Inquiries." The purpose was to data-ify inquiry trends and consider improvements that would reduce the burden on staff.

■Overview and Results of the Demonstration Experiment

In this demonstration experiment, conducted from January to February 2026, call recording data (approximately 1,800 cases) of staff responses to inquiries at the main business office telephone were collected, transcribed into text, and organized into a structured knowledge database. As a result, the accumulated knowledge (approximately 270 cases) was confirmed to contribute to the accurate identification of inquiry purposes and the smooth referral to appropriate departments. This is expected to be utilized for delivering accurate information to citizens without making them wait, and for verifying the feasibility of future AI-based automated responses.

Furthermore, the analysis visualized the process of identifying the purpose from various inquiries, specifically revealing concentrated patterns of routine inquiries related to My Number Cards and tax returns during specific periods. This confirmed the existence of many cases that can be resolved without staff intervention and provided a concrete path towards future verification of AI-based automated responses and chatbot implementation, as well as reducing the burden on staff.

■Persol Business Process Design's Strengths in Knowledge Generation

In knowledge generation for inquiry handling, inquiry content and knowledge do not always have a one-to-one correspondence. For the many patterns of inquiry content, operators respond from various angles to connect the inquirer with the knowledge they truly seek. Our strength lies in being able to build a database that includes the knowledge possessed by these operators.

■Future Outlook

Based on the results of this demonstration experiment, Persol Business Process Design will proceed with building a standardization and efficiency model for main telephone responses, and with organizing tasks that require human intervention versus those that citizens can resolve themselves through FAQs and automated responses. Furthermore, we will support the development of this initiative into telephone inquiry services within ward offices, and the long-term administrative DX measures such as the introduction of AI-based automated responses and chatbots. Utilizing the insights gained from this demonstration, we will also strive to support the efficiency and quality improvement of counter operations in local governments nationwide.