VICTOR CONSULTING Develops “Genba Capability Enhancement System 8/2,” a Generative AI Support Tool for Manufacturing Sites

VICTOR CONSULTING, based in Nagoya, Aichi Prefecture and led by Mikio Katsu, has developed “Genba Capability Enhancement System 8/2,” a generative AI-powered practical support system for small and medium-sized manufacturers and businesses with on-site operations. The system supports the creation of four types of documents: work procedure manuals, man-machine charts, cause-and-effect diagrams, and improvement proposal sheets. It is designed to help organize on-site work, time usage, causes of problems, and improvement ideas in a short period of time, turning them into materials that can be used for workplace improvement, new employee training, and skills transfer. The name “8/2” represents the concept that AI creates 80% of the structure, while humans contribute the remaining 20% of on-site knowledge. Rather than leaving workplace improvement entirely to AI, the system has AI provide standard templates and drafts, while humans add site-specific observations, experience, and judgment to complete documents that can be used in actual operations. The system was developed in response to a common challenge in small and medium-sized manufacturing companies: shop floors have valuable knowledge, but not enough time to document it. Many sites face labor shortages and reliance on veteran workers, while struggling to allocate enough time to standardize daily operations or conduct improvement activities. Examples include work procedures being explained verbally, new employee training being left to the shop floor, lack of visibility into how people and machines spend time, insufficient analysis of defects and rework, and improvement ideas ending as verbal suggestions without becoming formal proposals. To address this, VICTOR CONSULTING built the system in-house using generative AI, enabling users to create the forms needed for workplace improvement quickly. AI handles standard structuring tasks such as suggesting general procedures, organizing factors, drafting forms, writing text, visualizing information, and preparing proposal outlines. Humans add the site-specific judgment that only they can provide, such as subtle abnormalities, operator experience, equipment and process conditions, exceptions, reproduction conditions, feasibility, costs, time, and photos. A key feature of the system is its design that minimizes text entry. Instead of requiring users to write long documents from scratch, AI or templates first present standard procedures, factors, and form structures. Users then select, adjust, reorder, add photos, and supplement figures as needed to create practical documents for real work. The system consists of four main functions. First, work procedure manual creation: it provides process templates for machining, NC lathe processing, pressing, deburring, visual inspection, dimensional measurement, assembly, cleaning, packaging, and daily inspections. Users select a process category, load standard procedures, remove steps that do not fit their site, add necessary steps, and change the order using buttons. Each step can include photos, key points, precautions, and check items. The output is an Excel file containing a work procedure manual, photo collection, and input data sheets. Second, man-machine chart creation: users can record human work, monitoring and waiting, and machine operation and stoppage using smartphone buttons. Standard mode records activity with two buttons: “Start Work” and “Start Monitoring/Waiting.” Detailed input mode allows users to set site-specific buttons such as part loading, part removal, jig setup, next-part preparation, visual inspection, deburring, and measurement. Results are displayed as interval logs, summaries, and simple man-machine charts, and can also be exported to Excel sheets named log, summary, and events. Third, cause-and-effect diagram creation: when users enter a problem they want to solve, AI organizes general factors based on the 4M framework of Man, Machine, Material, and Method, then creates a draft cause-and-effect diagram. However, the purpose of this application is not to draw a neat diagram. It is to have AI organize 80% of the general factors, then use diagnostic questioning to draw out the subtle observations, differences, exceptions, and reproduction conditions known only to people at the site. AI asks questions based on comparison, paradox, and analogy to extract information that helps approach the root cause. Through repeated questioning, the system creates root cause candidates, next items to verify, a root cause exploration report, and text and JSON that can be passed to the improvement proposal creation AI. Fourth, improvement proposal creation: the system reads linked text or JSON output from the cause-and-effect diagram function and creates an improvement proposal. Before drafting the proposal, AI asks one final confirmation question to clarify feasibility, points of concern for the recipient, and constraints. Based on the answer, the system outputs a detailed proposal in Markdown format on screen, then compresses it into a one-page business improvement proposal in Excel. The Excel proposal can include the project name, person in charge, current problems, causes and root cause candidates, required costs and preparations, improvement details, expected effects, methods for verifying results, current-state photos, proposed or post-improvement image photos, and quantitative information such as costs, time, and number of cases. VICTOR CONSULTING says the system is not intended to be sold as a finished packaged software product. It was built as a practical support tool to be improved while being used in actual manufacturing support settings. Because it was developed in-house using generative AI, the company emphasizes its ability to incorporate feedback from the field and reflect fixes or feature additions in a short time. The initiative also serves as a practical example showing that people with shop-floor experience can use generative AI to build business support tools tailored to their own companies. The main target users include managers of small and medium-sized manufacturers, plant managers and manufacturing supervisors, on-site leaders promoting improvement activities, business support staff at credit unions, chambers of commerce, support organizations, and specialists. While manufacturing is the primary target, the system can also be applied to logistics, warehousing, office work, services, medical care, and nursing care, as long as the work involves people, operations, mistakes, waiting, variation, and rework. Mikio Katsu, representative of VICTOR CONSULTING, spent many years at Mitsubishi Heavy Industries working in manufacturing sites and production management, and served as a manufacturing department manager before becoming independent as a certified SME management consultant. He said manufacturing sites contain many seeds for improvement, but activities often stop at the stage of turning them into procedure manuals, visualizing time usage, digging into causes, or compiling proposal documents. In his view, generative AI does not replace on-site judgment; rather, it supports drafting and structuring so that people at the site can focus on what they truly need to think about. Going forward, VICTOR CONSULTING plans to expand the system into support services including workplace improvement using Genba Capability Enhancement System 8/2, creation support for procedure manuals, cause-and-effect diagrams, and improvement proposals, use of man-machine charts under labor shortages, and support for companies building their own internal business applications with generative AI. The system is an independent initiative by VICTOR CONSULTING and is unrelated to Katsu’s former employer. It does not hand workplace improvement over to AI, but supports the judgment of site personnel and managers. Outputs are intended to be used after on-site verification and review by relevant stakeholders. Company overview: VICTOR CONSULTING, representative Mikio Katsu, based in Nagoya, Aichi Prefecture. Business areas include management consulting for small and medium-sized manufacturers, workplace improvement support, generative AI utilization support, cash-flow and business improvement support, and services as a certified management innovation support organization. Website: https://victorconsulting.jp/. Details: https://victorconsulting.jp/genba-ryoku-8-2/. Contact: https://victorconsulting.jp/contact/.