Tokyu Construction Co., Ltd. (Headquarters: Shibuya-ku, Tokyo; President: Mitsuhiro Terada; hereinafter "Tokyu Construction") and NTT Sonority Co., Ltd. (Headquarters: Shinjuku-ku, Tokyo; President: Hiroshi Sakai; hereinafter "NTT Sonority") have launched a joint field trial aimed at improving safety management standards and operational efficiency at construction sites through the use of voice AI.
In this field trial, NTT Sonority's newly provided voice AI solution will be introduced to construction sites to verify the efficiency of documentation tasks—such as morning assemblies, risk prediction activities, staff safety meetings, and coordination between partner companies—and to validate the advancement of safety management standards at construction sites.
This initiative marks the first instance of NTT Sonority's voice AI being utilized for safety activities at a construction site.
[Background and Purpose of the Initiative] The construction industry faces challenges in improving productivity due to an aging workforce, a shortage of young talent, and legal caps on overtime work.
Meanwhile, at construction sites, meetings and record-keeping to protect worker safety are critical processes that cannot be omitted from the perspective of accident prevention. During meetings, it is necessary to record the content of discussions; however, manual recording often leads to omissions, posing a risk that could lower the quality of safety management. Furthermore, a significant amount of time is spent on safety management in daily operations, leading to calls from the field for improved operational efficiency.
To address these issues, Tokyu Construction and NTT Sonority are focusing on the audio from meetings held at construction sites, utilizing voice AI to improve operational efficiency and safety management standards.
[Details of the Field Trial] (1) Utilization of Voice AI Solution provided by NTT Sonority NTT Sonority has contributed to productivity improvements in various fields, such as construction sites and customer service, through "Voice DX" using devices and software powered by sound technology. In this field trial, we will combine dedicated microphones equipped with noise-cancellation technology with new voice utilization solutions to verify high-precision audio recording in the high-noise environments typical of construction sites.
Recorded audio is automatically converted into text, and generative AI handles the entire process from summarization and evaluation to report creation. As the first step toward the social implementation of this solution—which converts natural "voices" from the field into insights and business reports—we will configure the system to be optimized for Tokyu Construction's field operations and verify the automatic generation of reports based on actual conditions, as well as the system's practicality.
(2) Target Operations for Verification In this field trial, verification will be conducted for the following four types of field operations:
[Table: Target Operations and Main Verification Content] - Morning Assembly: Automatic recording and summarization of statements, AI evaluation of safety instructions. - Risk Prediction Activities (KY): Automation of recording via voice, AI-based evaluation and scoring. - Staff Meetings: Automatic recording of meeting minutes, listing of action items. - Coordination between Partner Companies: Automatic organization of communicated items, presentation of past accident case studies.
[Expected Outcomes of the Field Trial] (1) Efficiency in Documentation Tasks We will confirm that it is possible to automate recording using voice for the four tasks: morning assemblies, risk prediction activities, staff meetings, and coordination between partner companies. This is expected to significantly streamline the recording, organizing, and reporting processes that were previously performed manually by staff.
(2) Improving Safety by Turning Unrecorded Information into Assets We will accumulate and utilize audio information from construction sites that was previously lost without being recorded as digital data. The accumulated data can be used for reviewing safety management and training young employees. Additionally, the AI's ability to present past accident case studies will allow even less experienced employees to provide specific safety instructions.
(3) Improvement of Safety Awareness Features that allow AI to evaluate and visualize the content of morning assemblies and risk prediction activities will contribute to improving the safety awareness of participants. Furthermore, the mechanism itself, where conversation content is automatically recorded and evaluated, is expected to encourage behavioral changes in site managers and foremen.
(4) Adaptation to High-Noise Environments Using dedicated microphones equipped with NTT Sonority's patented technology, we will confirm that stable audio recording is possible even in the high-noise environments unique to construction sites. This will also contribute to improving the accuracy of AI-generated reports.
[Future Policy] Tokyu Construction will verify the results of this field trial and begin the sequential introduction of the voice AI system provided by NTT Sonority to its construction sites.
Moving forward, we aim to digitize all processes in the safety construction cycle, while working on integration with external systems and further advancing generative AI. Additionally, we intend to fully commercialize the knowledge gained through this initiative to other operators outside the construction industry, aiming for deployment across a wide range of sectors, including general contractors, manufacturing production plants, plant-related facilities, and railway transportation.
By leading construction site reform through the use of voice AI, Tokyu Construction will contribute to the "realization of safe and secure construction sites" and the "improvement of the working environment in the construction industry."
FACT BOX
- Source: PR TIMES
- Category: research