Nitto Kogyo, working with IBM Japan, has developed multimodal generative AI that automatically extracts information needed to prepare quotations for distribution boards from equipment drawings, and will gradually begin operating it at 23 locations nationwide.
On 20 October 2026, Nitto Kogyo announced that, with support from IBM Japan, it had developed a web application that extracts information needed to prepare quotations for distribution boards through drawing analysis. The application will be used at 23 locations nationwide, and an operating environment is being established for approximately 50 sales representatives, with practical operations scheduled to begin sequentially from the same month. By incorporating generative AI into the core of its business processes rather than using it merely as a tool, the company aims to accelerate quotation work and standardize quality.
The initiative is driven by worsening labor shortages and the increasing reliance on individual expertise. Preparing quotations for distribution boards has required the interpretation of various drawing formats, regulations and product specifications. Competitors are also advancing their digitalization efforts, and Nitto Kogyo spent approximately 18 months working with IBM to organize quotation know-how and structure business knowledge.
The application developed this time is a web application that automatically extracts circuit information, equipment information and special notes using multimodal AI when building drawings are uploaded. It adopts the reference architecture of an integrated AI platform. It also implements the generative AI methodology "IBM Consulting Advantage" and an AI agent framework, giving it a configuration that can accommodate future technological advances and changes in business requirements.
Nitto Kogyo and IBM Japan are examining a company-wide AI utilization platform, including AI governance, and will establish a framework for using AI safely and effectively through operations at 23 locations nationwide. Going forward, they plan to improve drawing-reading accuracy and expand the scope of coverage based on feedback from frontline operations, while considering the rollout of quotation data and know-how to related business processes.