CrowdWorks Inc. will participate in a government-led initiative to convert the experience and work expertise of skilled manufacturing workers into AI training data. It will structure unstructured materials, including drawings and work procedures from refining and petrochemical operations, for use in developing AI models that support on-site workers.
CrowdWorks Inc. announced on the 10th that it would collaborate with the agency AI company インハンス and participate as a partner in the Ministry of Trade, Industry and Energy’s “FY2026 Manufacturing Tacit Knowledge Platform AI Model Development Project.” Tacit knowledge refers to the decision-making criteria and expertise accumulated by skilled workers in the field. The aim is to convert this knowledge into data that AI can use.
インハンス will oversee the project, developing a model that links skilled workers’ experience with on-site materials and provides relevant information when AI is asked about work procedures or knowledge. CrowdWorks Inc. will preprocess unstructured on-site data and convert it into text data for AI training.
CrowdWorks Inc. will deploy its proprietary preprocessing solution, the “Alpina Knowledge Compiler,” to automatically structure elements containing technical information, such as tables, graphs, and formulas in documents. It plans to build an environment capable of processing up to 7.5 million pages of materials during the project period.
The processed data will be used to build an ontology and develop AI guide models that support on-site workers. Going forward, the company plans to expand its existing AI training data business into the manufacturing AI transformation (AX) field, with the conversion of know-how contained in on-site documents into high-quality data identified as the core focus.