Enhans, an enterprise agentic AI company, said on Tuesday it will carry out a project with GS Caltex to develop an AI model based on tacit manufacturing know-how that links skilled workers’ experience and judgement criteria at production sites to data.

The research and development is part of the Ministry of Trade, Industry and Energy’s "2026 AI model development project based on manufacturing tacit knowledge". Enhans will serve as the lead research and development institution and oversee overall development. GS Caltex will participate as the demand company, with OpenGraphLabs as a partner for parsing unstructured documents, and Chonnam National University and GIST participating as joint research institutions.

Through the work, Enhans will connect GS Caltex equipment-related structured data, unstructured documents and maintenance history with judgement criteria and exception-handling experience secured through interviews with on-site engineers. It plans to structure relationships such as equipment, parts, abnormal signs, causes and measures based on 832,996 GS Caltex facilities and about 2,900 unstructured documents.

It will identify, among engineers in the process, mechanical, electrical and instrumentation fields, the sequence for checking equipment conditions, criteria for narrowing down causes and additional considerations in exceptional situations, and reflect them in the ontology design.

The knowledge built will be implemented as a guide chatbot for on-site use. If on-site staff ask questions in natural language about equipment abnormalities or maintenance situations, the system will present candidate causes and inspection sequences. It will also allow them to check original sources that can serve as grounds for judgement, such as past maintenance history and technical standards, the company said.

Enhans Chief Executive Lee Seung-hyun (이승현) said it was important to go beyond leaving skilled workers’ experience simply in documents and convert it into knowledge that companies can use continuously. He said the company will build a system that supports on-site staff in checking the information and grounds for judgement they need and resolving problems more efficiently.